You are on page 1of 11

Energy and Buildings, 11 (1988) 103 - 113

103

Street Design and Urban Canopy Layer Climate


T. R. OKE

Department of Geography, The University of British Columbia, Vancouver, B.C. V6T 1W5 (Canada)

SUMMARY

Planning is always involved in making


choices between alternatives. In the case o f
designing for street climate the objectives may
be mutually exclusive. For example, whilst
open geometry is conducive to air pollution
dispersion and solar access, a more densely
clustered arrangement is favourable ~or shelter
and energy conservation.
This dilemma is investigated by reviewing
the results o f recent urban canyon field
studies and o f scale and mathematical modelling. By concentrating on quantifiable relations it appears that it may be possible to find
a range o f canyon geometries that are compatible with the apparently conflicting design
objectives o f mid-latitude cities. I f this is
correct, traditional European urban forms are
climatically more favourable than more modern, especially North American, ones.
THE PROBLEM

In a previous paper we have argued for


urban climatology to become a more predictive science so that its findings can be of
direct value in urban planning and design [1].
The present paper seeks to show by example
how this can be achieved. It poses the simple
but very fundamental question "Does urban
climate research have quantitative guidelines
to offer regarding street geometry?" It attempts to provide some answers by reviewing
relevant research findings and interpreting
them in terms that are readily understood,
and easily controlled, by design professionals,
viz., street dimensions and building density.
When starting to consider such a question
it is easy to become overwhelmed with the
vast range of possibilities and special cases.
These are associated with the almost infinite
combination of different climatic contexts,
urban geometries, climate variables and design
0378-7788/88/$3.50

objectives. Obviously there is no single solution, i.e., there is no universally optimum


geometry. However, this should not stop us
seeking general guidelines as long as they are
flexible enough to cater to special needs and
situations. We certainly do not want a rigid
'solution' whose blind application leads to
further problems.
The essence of planning and design is making choices between alternatives. This is not
an easy task, especially when it involves social
and economic values. Even here, where we are
mainly concerned with physical outcomes,
the choices are not obvious. We can illustrate
this by considering four of the most basic
objectives confronting those charged with
designing for street climate. For a mid- or
high-latitude city these goals may be:
(1) to maximize shelter. The obvious needs
are to ensure the safety and comfort of pedestrians by not exposing them to high winds
which might impede their progress or even
blow them over, or lead to high wind chill or
driving rain. Shelter may also avoid disruptive
snowdrifting and contribute to energy conservation by reducing turbulent heat losses.
(2) to maximize dispersion of pollutants.
In the case of streets the most important
source of pollution is probably that from
vehicle traffic on the canyon floor and the
primary concern is to minimize negative
impacts on receptors such as people (pedestrians, occupants of vehicles and houses) or
vegetation.
(3) to maximize urban warmth. By creating
a street structure which enhances the urban
heat island effect the climate will help reduce
pedestrian discomfort and the need for space
heating in the buildings. These benefits will be
most obvious at night and/or in the cold season. It is assumed that heat stress is not a
major problem in such climates.
(4) to maximize solar access. In order to
make best use of solar energy, either passively
Elsevier Sequoia/Printed in The Netherlands

104
or via active collectors, the geometries associated with beam radiation and the urban
structure must be favoumble. Similarly shading of canyons must not be so great as to
hinder daylighting of building interiors or to
cast a general gloom within canyons.
The objectives and the structures they dictate are in conflict. The firsttwo goals require
opposite structures; shelter is best provided
by narrow streets and compactness, whereas
dispersion demands separation and low building density. Similar dichotomies exist between
the other two objectives: warmth is promoted
by compactness but access by openness. Thus
full compatibility is possible only if the designer is concerned to provide either shelter
and warmth or dispersion and access. In general this happenstance is unlikely to occur.
A city is more likely to wish to foster a mix
of objectives. [As an aside we should note
that if some objective means of ranking these
goals were available, such as a cost-benefit
analysis or hazard assessment, it would greatly
help in the overall evaluation and selection of
the compromise structure.]
In the following we will review our knowledge of relationships between urban geometry
and street climate as they relate to each of the
preceding four goals. Then, assuming most
cities seek to meet each of these goals at least
minimally, we will investigate whether there
is a 'zone of compatibility' (i.e.,a range of
canyon geometries and building densities
which avoid the worst aspects of not providing shelter,dispersion, warmth or access).
Our consideration is restricted to the following limited example:
--a mid-latitude (approximately 45 ) city
where winter space heating is necessary,
vehicle traffic is sufficientto provide a significant source of pollutants, and the weather
includes wide ranges of wind speed and solar
irradiance.
--the street canyon is the basic geometric
unit. It can be approximated reasonably by a
two<limensional cross-section (i.e., we will
neglect street junctions and assume the buildings flanking the canyon are semi-infinite in
length.
-- the urban cross-section is approximated by
a simple repetition of these street canyon units.
- - t h e predominant direction of airflow is
approximately normal (say +30) to the long
axis of the street canyon.

Most of these restrictions can be relaxed or


changed to accommodate other circumstances
such as different latitudes, climates, cubic
structures, winds parallel to the street, etc.
Given the configuration we have specified
we can limit the geometric descriptors to two
simple measures. These are the ratio H/W,
where H is the average height of the canyon
walls and W is the canyon width (see Fig. 1);
and the building density, ~ = At~A1 where Ar
is the plan or roof area of the average building
and A1 is the 'lot' area or unit ground area
occupied by each building (see Fig. 3). H/W
is sometimes referred to as the aspect ratio.
These geometric measures do not explicitly
include compass orientation although this
will be mentioned in relation to solar radiation. Any final design 'solution' regarding
the layout of the street pattern in relation to
urban canopy layer climate must also address
this factor.

SHELTER AND URBAN GEOMETRY


The wind climate around an isolated obstacle such as a building is well documented
[2, 3]. The flow in an envelope surrounding
the building is perturbed. There are three
main zones of disturbance: ahead there is a
bolster eddy vortex due to flow down the
windward face, behind there is a lee eddy
drawn into the cavity of low pressure due to
flow separation from the sharp edges of the
building top and sides, and further downstream is the building wake characterized by
increased turbulence but lower horizontal
speeds than the undisturbed flow. Only the
first two zones are illustrated in Fig. l(a).
The flow over arrays of buildings is less
well understood. If the buildings are well
apart (H/W > 0.05) their flow fields do not
interact. At closer spacings, such as that in
Fig. l(a), the wakes are disturbed. When the
height, spacing and density of the array
combine to disturb the bolster and cavity
eddies, this isolated roughness flow regime
changes to one referred to as wake interference flow (Fig. l(b)). This is characterized by
secondary flows in the canyon space where
the downward flow of the cavity eddy is reinforced by deflection down the windward
face of the next building downstream. At
even greater H/W and density, a stable circula-

105

(a) Isolated roughness flow

;
I~

(/3) Wake interference flow

=I

(C) Skimming flow


l

",~ ~ s , ~ ' ~ ~ ~ ' ~ , . . . w

Fig. 1. The flow regimes aseochited with air flow over building arrays of Lncreasing

tory vortex is established in the canyon and


transition to a skimming flow regime occurs
where the b u ~ o f the flow d o e s n o t enter the
canyon (Fig. 1(c)). The transitions between
these three regimes occur at critical combinations o f H/W and L/W (where L is the length
of the building norm~d to the flow) as given
in Fig. 2.
If we assume that these wind tunnel results
apply to arrays of buildings in cities, we have
the basis for a f~st~)rder view of the effect o f
canyon geometry and building density on
shelter. We d o not, however, have a full re~tionship b e t w e e n wind speed reduction and
these measures of geometry. To give some
~tea we can n o t e that Nakamura and Oke [5]
suggest the simple linear form:

Ucanyon ----Pfiroof
where ~ is the mean horizontal wind speed
and p is a diminution factor which depends
on H/W and the measurement levels. They
show for wind speeds up to 5 m s-1, with
H/W ~ 1, and canyon centre and above-roof
measurement at heights o f a b o u t 0 . 0 6 / / a n d
1.2H respectively, that p ~ 2/3. Presumably
at smaller H/W, p approaches unity and
shelter is lost However, we should reiterate
that our concern is with the comfort and
safety o f pedestrians and heat loss from building walk. Both tend to be concentrated at the
sides o f the canyon, where we can anticipate
greater shelter than at the canyon centre.
There appears to be little useful emp~ical
information available on this point. Sh~il~rly,

H/W.

0.20

0.25

~H

Isolated roughness flow

.222;-

0.33

0.50

i
1
2
4

Skimming
Cube
I
1

1
I
2

I
3

I
4

I
5

I
6

Canvon- ~
I
7

L/H

Fig. 2. Threshold lines dividing flow into three regimes as functions of the building (L/H) and canyon
(H/W), geometry. Modified from a diagram in ref. 3
based on the wind tunnel restflts of ref. 4.

there is little knowledge on which to base


judgements regarding turbulence conditions
across a street. Gustiness is as important as
mean wind speed for m a n y applications [6].
It i~ very difficult to choose an objective
criterion for the minimum acceptable a m o u n t
of shelter. The criterion, or more probably
criterm, should be based on the objectwe(s)
involved. If concern i~ for pedestrmn comfort
and safety, rehitionships are avflable to calcu~ t e the e f f e c t of winds on the thermal corn,
fort and mechanical buffeting o f persons,
e . g . [ 7 - 9]. These, together with a knowledge
o f the general wind and temperature climatology and the effects of geometry, could be
used t o set the threshold o f acceptable conditions at a given location. The most pressing
need is to more accurately and fully establish

106
wind and turbulence diminution factors as a
function of H/W and position within the
canyon. This could be achieved via field observation or models (e.g., wind tunnel [2 - 4,
6] or numerical [10] models). Before use it
is essential that such models have been validated using field data.
Given the present state of uncertainty it
seems prudent to suggest no more than a very
general threshold based on Fig. 2. We presume
that some shelter is provided by wake interference and that the transition to skimming
flow bestows significant shelter. Using Fig. 2
for L / H values typical of canyons, we suggest
that H/W ~ 0.65 should ensure considerable
protection and, therefore, a minimum acceptable value may be somewhere in the middle of
the wake interference regime at about 0.4.
This minimum is based on fully open canyons;
obviously if additional protection such as
trees or shop window awnings are provided
this limit becomes more conservative.

DISPERSION AND URBAN GEOMETRY

Urban geometry affects the capacity to


disperse pollutants in a city on at least two
scales. The total array of roughness elements
affects the production of mechanical turbulence, the form of the vertical wind profile
and the depth of the urban mixing layer.
These are local or meso'scale effects. The
wake shed by each building and the circulation and turbulence associated with street
canyons also produce micro-scale effects in
amongst, and just above the buildings. The
relevance of the latter scale to the present
discussion is obvious, that of the former perhaps needs elaboration. It is beneficial to
canopy-layer air quality to have the lowest
possible concentration in the boundary layer
a b o v e for two reasons. Firstly, this provides a
relatively clean source of air to be entrained
down into the canyons, and secondly, the
lower the upper-level concentration, the
greater is the vertical concentration gradient
for upward turbulent diffusion.
The critical surface parameter governing
the production of turbulence in the urban
boundary layer is t h e roughness length (z0).
The larger the value of z0 the greater is the
intensity of turbulence [11, 12] and the
greater is the depth of frictional influence.

The roughness length is related to t h e surface


geometry in the following way. As large roughness elements (here buildings and trees) are
added to an area of less background roughness, the total roughness is increased. The
taller the elements, and the greater their number, the greater is the resulting value of z0 up
to a certain density. Beyond this the addition
of elements serves to reduce z0. This is because interference between individual wakes
serves to smother their role in producing turbulence. The progression is related to that
shown in the (a) - (c) sequence of flow regimes in Fig. 1.
Both f i e l d a n d wind tunnel studies confirm
that the relative roughness length, zo/h (where
h is the average height of the elements),
increases with element density to a single
peak and then declines [4, 13]. In these
studies the element density is described using
either the roughness density (~) or the building density (~') -- see Fig. 3. For building-like
elements the peak of zo/h is typically 0.2 - 0.3
and occurs at densities of about 0.25 (range
0.13 - 0.32) [4, 12 - 14]. The maximum
roughness effect may approximately correspond to the transition between wake interference and skimming flow [14] which for
canyons implies H/W ~ 0.65.
Within the canopy layer, pollutant transport and diffusion is strongly dependent
upon the above-roof wind and turbulence and
the urban geometry [15]. For small H/W,
dispersion is good and is closely linked to the
horizontal wind speed. Although the spatial
distribution of pollution may show localized
areas of high concentration (e.g., in the lee
cavity of isolated buildings) there is good
Ar
As

Fig. 3. D i m e n s i o n s o f an average urban roughness


element (e.g., a building) and its lot, where h is its
height, A s is its 'silhouette' area (the vertical area o f
the element 'seen' by t h e approaching wind), A l i s its
lot area (unit ground area occupied by each element),
and Ar is the element or roof area. These measures
can be used to define a roughness density )~ = As/A1,
and a building density ~ = ArIA I.

107
exchange between ground-level sources and
the cleaner air in the mixing layer above. This
is not the case for H / W beyond the threshold
for skimming flow. The development of the
stable cross~anyon vortex circulation is associated with reduced exchange between the
canopy and boundary layers. The vortex is
tangentially driven by the above-roof flow
but the coupling becomes less effective as
H/W is increased.
Field studies show the vortex is associated
with pollutant concentration differences within the canyon, especially near the floor [16 18]. Flow d o w n the windward wall of the
canyon has relatively clean air near the top
but becomes more contaminated as it entrains
canyon pollutants towards the floor. The
return flow across the canyon floor traverses
the source of vehicular pollution. Concentrations are highest at the base of the leeward
wall and decrease with height. In deep canyons
a secondary vortex m a y develop giving two vortices rotating in opposite directions [18]. The
cross-street concentrations are then the reverse o f that with the single vortex, and the
base of the canyon becomes more heavily
polluted because the 'handover' of materials
between the two vortices is ineffective. With
weak above-roof winds the vortex does n o t
form, the c a n y o n air m a y become decoupled
from that above and stagnation occurs. This
has potentially serious air<luality consequences near the floor unless thermal differences are capable o f generating a circulation.
Quantitative estimates o f the rate of removal
of pollutants from the c a n y o n and their
replacement with air from above is not yet
possible [3, 18]. There is evidence to show
that the vertical vortices shed from the
corners of building-faces at intersections, and
the turbulence generated from a few taller
than average structures scattered within more
uniform arrays of buildings, are capable of
enhancing below- and above-roof exchange
to the advantage of canyon air quality [3,

19, 20].
From the foregoing it seems that the onset
of skimming flow roughly coincides with a
marked reduction in the dispersive capacity
of c a n y o n air. F r o m Fig. 2 this suggests an
H/W limit o f about 0.65. There is also some
indication that the Vortex circulation is
strongest with H/W ~ 1.0 [21]. Here we opt
for the more conservative value of 0.65.

It is very difficult, and probably not prudent, to set a rigid criterion for the canyon
geometry which provides the m a x i m u m acceptable degradation of air quality (or minim u m acceptable canyon dispersion). Presumably air quality standards provide the thresholds compatible with a given jurisdiction's
quality-of-life objectives, but in a given canyon
the chances of complying with the standards
depends on the strength of emissions and
climate as well as the geometry. Here again,
the development of physical models to simulate the likely concentrations arising from
different emission scenarios should be encouraged.
Thus, in summary, it is interesting to note
that H / W ~ 0.65 with a building density of
~0.25 m a y fortuitously provide both a maxim u m roughness effect for above-roof urban
airflow and an upper limit to satisfactory
dispersion from street canyons.
URBAN WARMTH

AND GEOMETRY

A compact city form promotes urban


warmth (i.e.,the urban heat island), especially
at night. In mid-latitude areas this is generany
regarded as favourable in the cold season
because it reduces the need for space heating
in houses and thereby promotes energy conservation. It should also provide a measure of
pedestrian comfort, benefits to plant and
animal systems and enhanced pollution dispersion (via thermal turbulence, reduced
stability and thermal breezes). These probably
outweigh such disadvantages as the added
heat stress in the summer, pollutant fumigation and increased chemical weathering, but
here again a full cost-benefit analysis would
be very instructive.
Three of the most obvious causes of the
canopy layer heat island that are governed
by urban geometry are the increased absorption of solar radiation caused by multiple
reflection, the reduction of turbulent sensible
heat transferout of the canyons due to shelter
and the reduction of long-wave radiation loss
from within the canyons due to the screening
by the flanking buildings. (For a full list of
heat island causes see ref.22.) W e have already
covered the relationship between geometry
and shelter, and the following section deals
with that between geometry and solarradiation
receipt; here we consider long-wave radiation.

108
The long-wave radiation balance (L*) of
most natural surfaces is negative, i.e., a net
heat loss. This is especially true for surfaces
open to the sky because the sky vault temperature, and often emissivity, are lower than the
corresponding values o f terrestrial surfaces
[23]. Hence the sky is a very important net
energy sink in the infra-red region of the
electro-magnetic spectrum. Therefore the
extent to which any point on a surface is
open to the sky is a crucial consideration in
its ability to cool. The sky view factor (~s) is
a dimensionless parameter used to express
this. For a horizontal surface with an unobstructed horizon ~ = 1.0 and the overlying
hemisphere it subtends is completely open
to the sky. Lesser values indicate the fraction
of the sky that remains in the view of the surface, and (1 -- ~s) is the fraction occupied b y
screening obstacles. Since these obstacles are
often terrestrial objects their radiative emission is greater than that portion of the sky
they obscure. Hence the long-wave emission
from the overlying hemisphere is enhanced,
thereby reducing the net radiative loss of the
surface.
The sky view factor can be directly related
to the H/W ratio for the simple canyons we
are discussing. Consider the radiation geometry for a point at the centre o f the floor o f
the canyon cross-section given in Fig. 4. If we
assume to a first approximation that the street
is infinitely long, then the view factor o f each
wan
is
~w

(1 - - c o s 0)12

where 8 = tan -1 (H/O.5W). Therefore the sky


view factor is
~ s = (1 - - ( ~ w , + ~ w , ) )

For canyons with symmetric cross-section,


even more simply, ~s = cos 6, and the approximate relation between the measures is:
H/W

~s

1.00 0.89

0.25
27

0.5
45

1.0
63

2.0
76

3.0
81

0.71

0.45

0.24 0.16

4.0
83

0.12

Similar calculations can be conducted for any


point on the canyon floor or walls, each of
which has its o w n radiation geometry. Figure
5 shows that for a given time at night the
value of L* at any point in the canyon is almost linearly related to its ~s value. Further

7/7/2.

v////////////////////////////~
1~
0~
',=

~-',

Fig. 4. Geometry of an unsymmetric canyon flanked


by buildings 1 and 2.
--90

-80
0 West wall
East wall
0 Canyon floor
Canyon top

-70
-60
E -50

a
, -40
..J
-30

_@o

-20
-10

'

0 i2

0.4

06
I

'

0 I8

1.0

Fig. 5. Observed values of net long-wave radiation


(L*) and calculated values of sky view factor (~/s)at
locations within an urban canyon. Average results for
two nights with cloudless skies and light wind (after
Nunez [24]).

we seen that points near the t o p of the canyon


that are more open to the sky are losing heat
(cooling) at a b o u t four times the rate near the
floor. By extension we can say that ceteris
paribus canyons with large H/W will cool
more slowly.
The m a x i m u m nocturnal heat island intensity (ATu-r(m,~)) has been shown to be linearly related to ~s in the city centre [25].
For convenience the same data have been replotted in Fig. 6 using H/W instead of ~,. The
results for more than 30 mid- (and a few high-)
.latitude cities and towns are well described by:
ATu.z(max) ffi 7.54 + 3.97

in(H/W)

Although this relation confirms the importance o f urban geometry to heat island intensity it does not conclusively demonstrate
which process is responsible. Since t h e maxim u m nocturnal value occurs with cloudless

109
i

12

10

0e

8
o

ooo.O+

<3
4

Europe
N. America
Australasia

H/W
Fig. 6. The empirical relationship between urban
canyon geometry (H/W in the city centre) and the
maximum heat island intensity (ATu-r(max)). Based
on data in Oke [25 ].u = urban, r = rural.

skies and calm the heat bR1ence is dominated


by the interplay between the net long-wave

energy drain and the heat storage in the urban


fabric, with turbulent influences being negligible. The agreement with the control suggested b y Fig. 5 is very suggestive that longwave screening is the key. With greater cloud
and wind t h e role o f this term should diminish, as does the heat island intensity. However,
it is also to be expected that H/W is related
to other physical features associated with the
density o f building development. Greater densities can be expected to be accompanied b y
greater amounts o f impermeable cover, materials with higher thermal admittance, and increased anthropogenic heat flux density. In
turn these features will favour sensible rather
than latent heat, heat storage rather than heat
exchange and extra energy availability.
It is very difficult to quantify the value of
the urban heat island in relation to energy
conservation or human comfort. There are
models to calculate energy loss from buildings
which include the importance o f site factors
such as the urban heat island [26]. The importance of b o t h wind (shelter) and temperature (heat island) can then be assessed in relation to the thermal characteristics of the buildings and the synoptic climatology. Empirical
studies show that it is possible to save 5 - 7.5%
of space heating costs per one Celsius degree
increase of m e a n daily temperature [27, 28].

Similarly the energy balance and bio-climate


of pedestrians can be modelled [29].
Again we face the question of choosing a
threshold. This time we need to k n o w the
geometry which will retain a desirable proportion of the heat island. If as a first-order assumption we accept that the form o f Fig. 6
only depends on the configuration o f urban
structures, we see that because of the logarithmic shape we can achieve considerable control
over the heat island with small changes in
canyon geometry at low values o f H/W. For
example, under ideal heat island conditions,
one third of the maximum possible intensity
is gained with H/W of 0.4, one half with 0.7
and t w o thirds with 1.0. We do n o t know the
ATu.r vs. H/W relationship for less than ideal
meteorological conditions b u t it seems reasonable to expect it to be of similar form. If
this is correct then we might arbitrarily say
that the minimum H/W acceptable is a b o u t
0.4, thereby maintaining a b o u t one third of
the heat island potential for a given city.

SOLAR RADIATION AND URBAN GEOMETRY


There are t w o aspects of direct relevance
under this heading: surface albedo and solar
access.
The total albedo of an urban system, and
therefore its ability to absorb solar radiation,
depends u p o n the albedo of the c o m p o n e n t
materials and their geometrical arrangement.
The importance o f geometry has been demonstrated using observations [24] and b o t h
numerical and scale models of canyon radiative exchange [30 - 34]. All approaches show
that for buildings of equal height the albedo
of a crenellated surface is lower than that of
a flat plane composed of the same materials.
It also appears that the effect increases with
latitude, and is more pronounced in the lowsun season [31, 33]. Further, it increases with
H/W [31 33] and is greater in E - W rather
than N - S oriented canyons [31]. Using different scale model arrays, Aida [33] found
absorption increments of 13 - 27% for H/W in
the range 0.5 - 2.0 when the flat plane albedo
was 0.40. He also suggested that this increment is mainly dependent on the plan density.
In a numerical simulation Aida and G o t o h
[34] studied the albedo of different canyon
arrangements as a function of the ratio WI/W 2
-

110
SLlfl

b~---W l

r-l~

W2---~

W1/W2
0.06
=

0.13
=

0.25
i

0.5
i

2
;

4
,

8
i

16
i

0.4

..,%......."~'
0.3

<

....

0.2
0

:~

~:

0.1

0,0-1.2

~----~

Min.
albedo

O. . . . .

0 = 60

.0

I
-01.9

-016

-01.3

0.0

0
8

0-----0

8 =

Z~ ......... ~

01.3

01.6

=20
= 40

=80

I
0.9

I
1.2

log (W 1/W 2)

Fig. 7. T h e d e p e n d e n c y o f t h e t o t a l s y s t e m a l b e d o
u p o n t h e s y s t e m g e o m e t r y as described b y t h e r a t i o
o f b l o c k - t o - c a n y o n w i d t h ( g r l / w 2 see t o p s c h e m a t i c ) .
Also includes t h e i n f l u e n c e o f solar z e n i t h angle (0).
Based o n n u m e r i c a l s i m u l a t i o n s w i t h H/W2 = 1.0.
A f t e r A i d a a n d G o t o h [ 34 ].

where W1 is the width o f the block elements


and W2 is the inter-block or street width (see
schematic in Fig. 7). Obviously W1/W2 is
proportional to W1/(W1 + W2), and therefore
~, for regular arrays. The interesting result
shown in Fig. 7 is that, when H/W is held constant at unity, there is a minimum albedo for
the system at log(W1/W2) = --0.3 to --0.6.
This means that in cases where we wish to
maximize solar absorption, such as mid-latitude cities in the cold season, the ideal W1/W2
is approximately 0.5. When translated into
blocks of semi-infinite length this gives ~
0.33.
Although minimizing the albedo increases
the absorptance of the total system for solar
radiation, it does n o t necessarily benefit positions near the canyon floor which m a y be in
shade. Indeed certain geometric arrangements
can create absorptance that is less than that
of a flat plane [31]. Here we are only considering the case of canyons with flanking build-

ings of approximately equal height, obviously


anomalously large structures can create energy deficient areas in their shadow.
The relevant limit for solar access to street
canyons depends upon the degree of penetration at the winter solstice. Penetration is
needed to facilitate solar energy gain b y
equator-facing walls, to provide sufficient
daylight for building interiors and to aid in
the comfort and psychological attitude of
pedestrians.
Table 1 gives some idea of penetration into
an E-W canyon b y indicating the a m o u n t of
an equator-facing wall that is open to directbeam irradiation at noon on the winter solstice. For such a canyon at a latitude of 45
with an H/W o f 1.0 we see that only the upper
39% is directly sunlit (Fig. 8(a)). An H/W o f
a b o u t 0.4 is required for the whole wall to be
directly lit (Fig. 8(b)). Full irradiation at the
solstice is probably t o o restrictive a requirement. A reasonable compromise m a y be an
H/W of 0.6 which would result in a b o u t t w o
thirds of the wall being lit.
TABLE 1
T h e area o f a n e q u a t o r - f a c i n g wall in a n E - W o r i e n t e d
s t r e e t c a n y o n t h a t p o t e n t i a l l y receives d i r e c t - b e a m
solar r a d i a t i o n at n o o n o n t h e w i n t e r solstice as funct i o n s o f l a t i t u d e a n d H/W. The s u n l i t area is expressed
as a p e r c e n t a g e f r a c t i o n o f t h e t o t a l wall area
Latitude

40
45
50

H/W
0.3

0.4

0.5

0.6

0.8

1.0

2.0

100
100
99

100
98
74

100
79
59

83
66
49

62
49
37

50
39
30

25
20
15

Since many mid-latitude cities experience


high frequencies of cloud in winter, the
amount of diffuse radiation for daylighting is
also important. The conventional criterion of
daylighting specialists is that an H/W o f 0.58
is appropriate at a latitude o f 45 [35, p. 74].
The corresponding ratios at 40 and 50 are
0.70 and 0.46, respectively.
Therefore we m a y tentatively conclude
that an H/W o f a b o u t 0.6 seems to be a suitable upper limit to maintain solar access in a
city at a latitude of 45 . Obviously there is
need to refine this conclusion on the basis o f
a full analysis of the passive solar energy gain
and illumination of the total urban system.

111

,. ,. ~ ~ ~ W i n t e r

~/////i/~j,i/////,

"/. S

(a)
s

~Winter

t',,'//'/'/,/jf/'/,//'//'/,/'/'///'/'/////////////////////

I~

~, I

(b)
Fig. 8. Angles o f i n c i d e n c e o f d l r e c t - b e a m solar radmt i o n at n o o n in a n E - W c a n y o n , in a city at 45N.
(a) I n a c a n y o n w i t h ./-//W = 1.0; ( b ) tt/W = 0.40.

This should include a recognition of the solar


radiation climatology o f the city in question
including the frequencies o f direct and diffuse
insolation. Similarly, the relative importance
of the effect of geometry on solar absorptance versus that o n solar access to the canyons cannot be simply decided, b u t one suspects that the solar access criterion is the
most critical.

A ZONE OF COMPATIBILITY ?

This paper demonstrates that a number o f


useful relationships exist b e t w e e n the geometry and the microclimate of urban street
canyons. They are potentially helpful to the
establishment of guidelines governing street
dimensions for use b y urban designers. The
relationships have the special merits o f being
quantitative and only depending on simple
measuI~ss.

Unfortunately there is less basis for linking


canyon climate characteristics to socioeconomic, comfort or safety objectives. In
m a n y cases the choice o f thresholds is arbitrary or largely subjective. There is a need to
refine these through further research b u t
there will always be an element o f value

judgement involved in setting the priorities


and acceptable limits in a given city.
D e s p i t e this problem we can summarize
the findings of the present exploratory study
and suggest some generalizations for our
hypothetical mid-latitude city.
(1) In terms o f H/W we find some degree
of agreement. A lower limit of a b o u t 0.4 is set
b y the need to provide some degree of shelter
and to retain a reasonable proportion o f the
heat island warmth. An upper limit of 0.60 0.65 ensures that b o t h atmospheric dispersion
and solar access is maintained within the street
canyons. Therefore, if all four goals are to be
satisfied, to at least a minimal extent, a compatible range of H/W is 0.4 - 0.6.
(2) There are rather few relationships available which use the building density, nevertheless it seems that roughness m a y be maximized
at ~ ~ 0.25 and absorptance at ~0.33. Tentatively we m a y suggest that a range of ~ between 0.20 and 0.40 is probably suitable.
If we apply the range of compatibility to
existing European and North American city
forms we find that neither conform in their
core areas for large cities ( > 1 0 s inhabitants).
In Europe the typical central area has H/W in
the range of 0.75 - 1.7, and in North America
it is 1.15 - 3.3 [25]. Given the lower values,
and the fact that the European data include
cities with very much larger populations, the
European results appear to be more favourable. Certainly the 'skyscraper' is a poor climatic form on the grounds of dispersion and
solar access near the ground. We might also
note that because such tall buildings j u t above
the general roof-level t h e y also create special
high-wind speed problems in localized areas
near their base, so t h e y do not provide uniform shelter either [6]. Although a survey of
H/W and ~" for European and North American
cities is n o t available, most observers would
agree that the compact form of residential/
suburban areas in the former are more likely
to conform with the suggested compatible
ranges than the sprawling suburbs of the latter.
Thus the m o d e m (especially) North American
city form comprised of t w o extremes -- a very
dense core and scattered suburbs -- is a poor
design climatically.
It is hoped that the reader o f this paper
will recognize the spirit of its intent. It seeks
to demonstrate b y example how urban climatologists m a y help in the formulation o f urban

112

design guidelines that are both quantitative


and easily understood. Clearly there are gaps
in knowledge and methods that must be
improved, and there is too much hand-waving
and speculation herein that must be refined
and verified. But if this paper encourages
others to ask simi]Ar simple questions regarding climatic design and to translate climatic
research into useful general guidelines it will
have achieved the author's aim.

10

11

12

]3
ACKNOWLEDGEMENTS
14

This work was supported by funds from


the Natural Sciences and Engineering Research
Council of Canada. The diagrams were drawn
by P. Jance.

15

16
REFERENCES

1 T. R. Oke, Towards a prescription for the greater


use of climatic principles in settlement planning,
Energy Build., 7 (1984) 1 - 10.
2 R. N. Meroney, Turbulent diffusion near buildings, in E. Plate (ed.), Engineering Meteorology,
Elsevier Science Publications, Amsterdam, 1982,
pp. 481 - 525.
3 R. P. Hosker, Jr., Flow around isolated structures
and building clusters: A review, A S H R A E Trans.,
91 (2B) (Feb.) (1985) 1671- 1692.
4 M. Hussain and B. E. Lee, A n Investigation of
Wind Forces on Three-Dimensional 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, Report
BS 56, Dept. Building Science, University of
Sheffield, 1980.
5 Y. Nakamura and T. R. Oke, Wind, temperature
and stability conditions in an E-W oriented urban
canyon, Atmos. Environ., submitted for publication.
6 A. D. Penwarden and A. F. E. Wise, Wind Environment Around Buildings, Building Research
Establishment Report, Dept. of the Environment,
HMSO, London, 1975.
7 A. D. Penwarden, Acceptable wind speeds in
towns, Build. 8ci., 8 (1973) 259 - 267.
8 T. V. Lawson, The Wind Environment of Build-

ings: a Logical Approach to the Establishment of


Criteria, Report No. TVL 7321, Dept. Aeronautical Engineering, University of Bristol, 1973.
9 A. G. Davenport, An approach to human comfort
criteria for environmental wind conditions, Proc.

CIB/WMO Colloquium Teaching the Teachers,

17

18

19

20

21

22

23
24

25

26

Swedish National Building Institute, Stockholm,


1972.
U. Sievers and W. G. Zdunkowski, A microseale
urban climate model, Contrib. Atmos. Phy&,
59 (1986) 13 - 40.
J. Counihan, Adiabatic atmospheric boundary
layers: A review and analysis of data from the
period 1880 - 1972, Atmos. Enviror~, 9 (1975)
871 - 905.
J. F. Clarke, J. K. S. Ching and J. M. Godowitch,
An Experimental Study of Turbulence in an
Urban Environment, Environmental Science
Research Laboratory, USEPA, Research Triangle
Park, 1982.
I. Seginer, Aerodynamic roughness of vegetated
surfaces, Bound.-Layer MeteoroL, 5 (1974)
383 - 393.
J. Counihan, Wind tunnel determination of the
roughness length as a function of the fetch and
the roughness density of three-dimensional roughness elements, Atmos. Environ., 5 (1971) 637 642.
R. A. McCormick, Air pollution in the locality of
buildings, Phil Trans. Roy. Soc. (London), A269
(1971) 515 - 526.
H.-W. Georgii, E. Busch and E. Weber, Investigation of the Temporal and Spatial Distribution of
the Emission Concentration .of Carbon Monoxide
in Frankfurt/Main, Report No. 11, Institute of
Meteorology and Geophysics, University of
Frankfurt]Main, 1967.
F. L. Ludwig and W. F. Dabberdt, Evaluation of
the APRAC-IA Urban Diffusion Model for Carbon Monoxide, Final Report CAPA-3-68 (1-69),
Stanford Research Institute, Menlo Park, 1972.
F. T. DePaul and C. M. Shieh, Measurements of
wind velocity in a street canyon, Atmos. Environs, 20 (1986) 455 - 459.
W. G. Hoydysh, R. A. Griffiths and Y. Ogawa, A
scale model study of the dispersion of pollution
in street canyons, ]>reprint No. 74-157, Proc. 6 7th
Annual Meeting, Air Pollution Control Assoc.,
1974.
J. B. Wedding, D. J. Lombardi and J. E. Cermak,
A wind tunnel study of gaseous pollutants in city
street canyons, J. Air Poll. Control Assoc., 27
(1977) 557 - 566.
I. Tanl, M. Iuchi and H. Komoda, Experimental
Investigation of Flow Associated with a Step or
Groove, Report No. 364, Aeronautical Research
Institute, University of Tokyo, 1961.
T. R. Oke, The energetic basis of the urban heat
island, Q. J. Roy. MeteoroL Soc., 108 (1982)
1 -24.
T. R. Oke, Boundary Layer Climates, Methuen
and Co. Ltd., London, 2nd edn., 1987.
M. Nunez, The energy balance of an urban canyon, Ph.D. Thesis, Dept. Geography, University
of British Columbia, Vancouver, 1974.
T. R. Oke, Canyon geometry and the nocturnal
urban heat island: Comparison of scale model
and field observations, J. ClimatoL, 1 (1981)
237 - 254.
R. Taesier, Urban climatological methods and
data, in T. R. Oke (ed.), Urban Climatology and

113

its Applications with Special Regard to Tropical


Areas, Report No. 652, WMO, Geneva, 1986,
pp. 199 - 236.
27 G. A. McKay and T. Allsopp, The role of climate
in affecting energy supply/demand, in W. Bach,
J. Pankrath and J. Williams (eds.), Interactions o f
Energy and Climate, D. Reidal Publishing Co.,
Dordrecht, Holland, 1980, pp. 53 - 72.
28 R. Taealer, Climate, Energy Demands and Energy

Consumption with Special Regard to Human


Settlement~, Report No. 2, Energy Special Applications Programmes, WMO, Geneva, 1981,
pp. 199 - 236.
29 G. Jendritzky and W. Niibler, A model analysing
the urban thermal environment in physiologically
significant terms, Archly. MeteoroL Geophy~
Bioklim., Ser. B, 29 (1981) 313 - 326.
30 C. D. Craig and W. P. Lowry, Reflections on the

31

32

33

urban albedo, Prepr. Conf on Urban Environ. and


Second Conf. on BiometeoroL, American Meteorological Society, Boston, 1972, pp. 159 - 164.
W. H. Terjung and S. S.-F. Louie, Solar radiation
and urban heat islands,Annals Assoc. Am. Geog.,
63 (1973) 181 - 207.
A. J. Arnfield, A n approach to the estimation of
the surface radiative properties and radiation
budgets of cities, Phys. Geog., 3 (1982) 97 - 122.
M. Aida, Urban albedo as a function of the urban
structure -- a model experiment, Bound.-Layer
MeteoroL, 23 (1982) 405 414.
M. Aida and K. Gotoh, Urban albedo as a function of the urban structure -- a two-dimensional
numerical simulation, Bound.-Layer MeteoroL, 23
(1982) 4 1 5 - 424.
M. Evans, Housing, Climate and Comfort, The
Architectural Press, London, 1980.
-

34

35

You might also like