Pergamon

PII:

Solar Energy Vol. 73, No. 2, pp. 105109, 2002

2002 Elsevier Science Ltd
S 0 0 3 8 0 9 2 X ( 0 2 ) 0 0 0 3 8 5 All rights reserved. Printed in Great Britain
0038-092X / 02 / $ - see front matter

www.elsevier.com / locate / solener

DAYLIGHT PREDICTION IN ATRIUM BUILDINGS

PAUL LITTLEFAIR
BRE, Garston, Watford WD25 9XX, UK
Received 17 August 2001; revised version accepted 28 March 2002

AbstractIt is often complex and difficult to predict the daylighting of an atrium building. The atrium is
usually expected not only to be an attractively daylit space in its own right, but also to deliver light to
adjoining spaces. Most of the light reaching these spaces will have been reflected within the atrium. This paper
outlines current guidance on daylighting design for atria and reviews published techniques to calculate daylight
within them, including a methodology to predict daylight factors within atria and the surrounding spaces.
2002 Elsevier Science Ltd. All rights reserved.

in the atrium. Measurements in real atria

(Littlefair and Aizlewood, 1998) suggest that
half or more of the light can be blocked. The
effective transmittance depends on the angle of
view; areas towards the top of the atrium tend
to be more affected than its base.
Getting light into spaces adjoining the atrium
can be difficult.
Solar shading of spaces adjoining the atrium
needs to be considered if glare is an issue, for
example in offices. Blinds or baffles will be
needed if sunlight can reach these spaces
through the atrium.
Effective control of electric lighting in an
atrium is essential if the full benefits of
daylight are to be realised. Because building
occupants do not expect to have control over
the lighting in an atrium, automatic (photoelectric) control is often best.

1. DAYLIGHT DESIGN IN ATRIUM

BUILDINGS

The atrium or enclosed courtyard forms a key

element in many recent building designs. One of
the perceived advantages of the atrium is that it is
seen as bringing natural light into the core of a
building.
A previous paper (Littlefair and Aizlewood,
1998) describes current guidance on daylighting
of atrium buildings. This can be summarised as
follows:
Atrium shape is a key parameter. A deep,
narrow atrium will be poorly lit. Splaying the
atrium, so that the top is wider than the
bottom, will give more daylight potential.
High reflectances in the atrium, particularly at
the higher levels, will boost interreflected light.
This can be achieved by reducing the glazing
or aperture area in the walls towards the top of
the atrium, or by using reflective glazing in
these areas.
Glazing areas need to be chosen with care to
admit enough daylight without excessive solar
heat gain.
Planting can be attractive, but many plants
require high illuminances (Baker et al., 1993).
To avoid excessive supplementary electric
lighting, it is best to choose those plants which
can survive at lower light levels.
Internal obstructions like walkways, cleaning
cradles, glazing bars and structural members
can significantly reduce the daylight available

2. PREDICTION TECHNIQUES

Tel.: 144-1923-664-874; fax: 144-1923-664-781; e-mail:

littlefairp@bre.co.uk

Predicting the distribution of daylight levels in

an atrium and its adjoining spaces can be difficult.
Three techniques can be used:
Model studies
Computer simulation
Analytical formulae
In scale model studies (Littlefair, 1993), it is
possible to assess the visual impact of the system
as well as measure the light admitted. To check
whether enough daylight reaches the interior and
that there are no gloomy areas, an overcast sky
study in an artificial sky is easiest. To assess solar
shading or glare protection, the real sun or a
heliodon can be used with a sundial, and assess-

105

106

P. Littlefair

ment made visually, coupled with internal luminance measurements if required.

For models used in daylight measurement, the
following guidelines need to be followed (in order
of importance):
(i) All the building surfaces must be present.
(ii) Access is needed inside the model for
measurement and visualisation.
(iii) Reflectances must be correct. Often models
are too light coloured which can be a big source
of error.
(iv) The model should be light-tight.
(v) External obstructions should be modelled
accurately, both in (scale) size and reflectance.
(vi) Window details such as atrium roof structure should be modelled if possible.
(vii) The scale of the model needs to be chosen
with care. A very big or very small model is hard
to test; 1:40 is a good compromise.
(viii) Plan measurement positions in advance
(not too many).
Some computer programs can calculate the
daylight factor at the base of an atrium. Most use
the point to point or radiosity algorithm. All
surfaces modelled are divided into a number of
elements, assumed perfectly diffuse. For simple
spaces with no significant areas of specular (mirror-like) reflection this is reasonably accurate and
quick to compute. To model complex spaces with
significant specular reflection ray tracing programs have been developed. One such program is
RADIANCE (Ward, 1994). This technique requires far more computer power than a radiosity
calculation, but in principle should be more
accurate since surfaces may be diffuse, specular
or of any semi-specular nature.
Measurements in an atrium model at BRE were
compared with RADIANCE predictions (Aizlewood et al., 1997). For the sky component at
the base of the atrium, there was complete
agreement between them. However, RADIANCE
underestimated the reflected light in deep, high
reflectance atria. Although the relative error is not
large, this is a matter of some concern because
there is no clear explanation for it (Aizlewood et
al., 1997).
The other way to estimate daylight in an atrium
is to use an analytical formula. This is considered
in detail in the next section.
3. FORMULAE TO PREDICT DAYLIGHT IN
ATRIA

Many authors have produced formulae or diagrams to predict the daylight factors within atria.

Usually these are based on scale model measurements or computer simulations. Aizlewood (1995)
has reviewed them in detail.
The British Standard Code of Practice for
daylighting (1992) uses the average daylight
factor DF av (Littlefair, 1988; Lynes, 1979) as a
way to judge a daylit space. The daylight factor is
the ratio of internal illuminance to external
horizontal unobstructed illuminance; DF av is the
average over the atrium base. If an atrium is to
look well daylit, then DF av should be 5% or more.
The average daylight factor on the base of the
atrium can be calculated using the following
formula (Littlefair, 1988):
WT g T fu
DF av 5 ]]]
A(1 2 R 2 )

(1)

The average daylight factor over all the atrium

surfaces (Lynes, 1979) is given by
WT g T fu
DF avs 5 ]]]
2A(1 2 R)

(2)

where W is the area of the atrium roof aperture

(m 2 ). T g is the diffuse visible transmittance of the
glazing, corrected for dirt on the glazing. Provision must be made for cleaning atrium glazing. T f
is a factor to allow for light blocked by the atrium
roof structure. u is the angle of visible sky in
degrees, measured as shown in Fig. 1. For a
completely unobstructed horizontal roof this
would give u as 180. In fact 200 would be a better
approximation here. A is the total area of the
atrium surfaces: roof, floor, walls and windows
(m 2 ). If the atrium sides are open, they should be
included in A by adding in the area of imaginary
atrium walls. R is the average reflectance of these
surfaces. For many atria R is fairly low. Even if
the atrium walls are white (reflectance 0.8), the
roof and side glazing (reflectance 0.15 single, 0.3
double) will reduce the average. If the atrium
sides are open, their reflectance is taken as 0.
Typical average values of R are 0.3 to 0.4.
These equations are most accurate for fairly
shallow atria (width at least half their height). If
an atrium has both roof and side glazing the
average daylight factor due to each is calculated
separately, then the two numbers are added
together.
The advantage of the average daylight factor is
that it can give a single number which summarises the overall daylit appearance of the atrium.
However, it does not give information about the
detailed distribution of light which is also important.

Daylight prediction in atrium buildings

107

Fig. 1. u is the angle of visible sky (in degrees) viewed in a vertical section from the centre of the atrium roof aperture.

The daylight factor on the walls of the atrium is

of particular interest because it is related to the
amount of light entering adjoining spaces. The
annex to this paper gives analytical formulae for
the daylight factor DF v on the vertical centre line
of a rectangular atrium.
The contribution from the atrium to the average
daylight factor DF sav in the adjoining space is
then given by (Littlefair and Aizlewood, 1998)
2A W T s DF v
DF sav 5 ]]]
A s (1 2 R 2s )

(3)

where A W is the net area of the glazing between

the space and the atrium (m 2 ). T s is the diffuse
visible transmittance of this glazing. For clean
clear single glass, a value of 0.8 can be used; for
double glazing 0.65; for no glazing 1.0. A s is the
2
total area (m ) of the room surfaces: ceiling, floor,
walls and windows, including those to the atrium.
R s is the average reflectance of the room. A
typical value is 0.5 for a light coloured space.
Often an adjoining room would also be lit from
the outside of the building. The average daylight
factors from each set of glazing can simply be
added together.
The average daylight factor measures the
amount of daylight in the adjoining space. However the distribution of this light is also very
important. Light from an atrium may not go very
far into an adjoining space. The depth to which
daylight penetrates into an adjoining space can be
roughly estimated using the no sky line (Littlefair,
1991). The no sky line divides those areas of the
working plane which can receive direct skylight,
from those which cannot receive any. Areas
beyond the no sky line will generally look gloomy
(Lynes, 1979), and supplementary electric lighting will be required. Fig. 2 shows typical no sky
lines in an atrium building.
The penetration of daylight into adjoining
spaces can be improved by:

Increasing the head height of the apertures

between the adjoining spaces and the atrium
Higher reflectances both in the atrium and
adjoining spaces to increase the amount of
reflected light
Innovative glazing systems (Littlefair, 1996),
like light scoops and shelves, can distribute
light from the atrium into adjoining spaces,
particularly with splayed atrium walls, where
the top floors are set back. They can also
decrease the contrast between the very bright
areas next to the atrium and the darker areas
further in.
Changing the roof profile to admit more light
from the side.

4. CONCLUSION

Predicting daylight levels within an atrium

building can sometimes be difficult or uncertain.
This paper has set out techniques to calculate
daylight factors on the atrium base and walls and
in adjoining spaces. The equations are based on
analytical calculations rather than empirical curve
fitting and hence should be applicable to a range
of atrium shapes and reflectances.

Fig. 2. No sky lines in an atrium building.

108

P. Littlefair

AcknowledgementsThis paper was produced as part of the

research programme of the Construction Directorate of the
Department of the Environment, Transport and the Regions,
United Kingdom. Thanks are due to Maurice Aizlewood for
his contributions to the work on which this paper is based.

APPENDIX A. DAYLIGHT FACTORS ON AN

ATRIUM WALL

Sky component on an atrium wall. This Appendix derives the daylight factor at a point P on the
vertical centre line on the atrium wall of a
rectangular atrium (Fig. 3). Assuming this point
cannot see any external obstructions through the
atrium roof, this is the sum of the sky component
and internally reflected component. Here we
assume that the whole of the horizontal roof is
glazed apart from a correction factor for glazing
bars and roof members.
Sky component (by Seshadris (1960) formula)
2.3
SC 5 ]]( b 2 b 9 cos g )
14p
2.2
1 ](sin b 2 sin b 9 cos 2 g ) 3 100%
7p
where
w
b 9 5 tan 21 ]]]
]]
2d 2 1 y 2
w
b 5 tan 21 ]
2y

w
sin b 5 ]]]]
]]]
2
2y 1 w 2 / 4
w
sin b 9 5 ]]]]]
]]]]]
2
d 1 y 2 1 w 2 / 4
y
cos g 5 ]]]
]]
2
d 1 y 2
by substitution:

3
w
y
w
SC 5 100 ? ] tan 21 ] 2 ]]
tan 21 ]]]
]]
]]
7p
2y
d 2 1 y2
2 d 2 1 y2

4
w
w ? y2
1 100 ? ] ]]]
2 ]]]]]]]
]]
]]]]
7p
4y 2 1 w 2 (4d 2 1 4y 2 1 w 2 ) ? (d 2 1 y 2 )

In a real atrium the sky component has to be

multiplied by T g and T f where T g is the glass
transmission and T f the transmission of the glazing bars.
The internally reflected component. To calculate the internally reflected component we divide
the atrium into two parts, above and below the
point P for which the daylight factor is required.
Then the internally reflected component is given
by
IRC 5 DF reflected from atrium base (below P)
1 DF reflected from atrium roof
1 DF reflected from walls in top part
(above P)
The first term in the equation is given by
DF reflected from base to P 5 DF b R b / 2
From Eq. (1)
WT g T fu
DF b 5 ]]]
A(1 2 R 2 )

Fig. 3. Dimensions of an atrium space (roof omitted for

clarity).

where W is the area of the atrium roof aperture

(m 2 ). T g is the diffuse visible transmittance of the
glazing, corrected for dirt on the glazing. Provision must be made for cleaning atrium glazing. T f
is a factor to allow for light blocked by the atrium
roof structure. u is the angle of visible sky in
degrees. (Fig. 1). A is the total area of the atrium
surfaces: roof, floor, walls and windows (m 2 ). R is
the area-weighted average reflectance of these
surfaces.
Throughout Appendix A we assume that if the
atrium walls contain open apertures into surrounding spaces, then they are included in any wall
areas and area-weighted reflectances (their reflectance being taken as zero).

Daylight prediction in atrium buildings

R b , the effective reflectance of the base of the

atrium, is given by the cavity reflectance formula
R b 5 R fwb A f /(A fw (1 2 R fwb )
where A f is the area of the floor, A fw is the total
area of floor and walls in the base of the atrium,
and R fwb is the area weighted average reflectance
for these surfaces (floor and walls).
The second term in the IRC equation is given
by
Daylight factor reflected from the atrium roof
5 DF c R c F

DF c 5 T g T f R a
with R a , the cavity reflectance of the entire
atrium, being
R b 5 R fw A f /(A fw (1 2 R fw )
where A f is the area of the floor, A fw is the total
area of floor and walls in the whole of the atrium,
and R fw is the area weighted average reflectance
for the floor and walls.
If the ceiling is a uniform diffuser, the factor F
is given by (Hopkinson et al., 1966)

WT g T fu
DF wt 5 ]]]]
2A t (1 2 R t )
In this equation A t is the area of the top surfaces
in the atrium (2wy 1 2yd 1 2wd). R t , the average
reflectance of the top part of the atrium includes
the cavity reflectance of the bottom part of the
atrium R b , thus
R t 5 (2 w d R c 1 2 (w 1 d)y R w 1 2 w d R b ) /A t
REFERENCES

where R c is the area weighted average reflectance

of the atrium roof, including glazing, and the
daylight factor on the atrium roof DF c is given by

1
w
y
w
F 5 ] tan 21 ] 2 ]]]
tan 21 ]]]
]]
]]
2
2
p
2y
d 1y
2 d 2 1 y2

109

Finally the daylight factor caused by light reflected from the top part of the walls is given by
Daylight factor from top walls
5 DF wt R w (0.5 2 F )
where R w is the area weighted wall reflectance.
The average daylight factor on the top walls is
given by (from Eq. (2) of the main paper)

Aizlewood M. E. (1995) The Daylighting of Atria: A critical

review. ASHRAE Conference, June 1995, San Diego.
Aizlewood M. E., Butt J. D., Isaac K. A. and Littlefair P. J.
(1997) Daylight in atria: a comparison of measurements,
theory and simulation. In Proceedings Lux Europa, Amsterdam.
Baker N., Fanchiotti A. and Steemers K. (1993). Daylighting
in Architecture, A European Reference Book, James and
James, London.
British Standards Institution (1992) Code of practice for
daylighting. British Standard BS8206, Part 2. BSI, London.
Hopkinson R. G., Petherbridge P. and Longmore J. (1966).
Daylighting, Heinemann, London.
Littlefair P. J. (1988) Average daylight factor: a simple basis
for daylight design. Building Research Establishment Information Paper 15 /88. Construction Research Communications, Garston, UK.
Littlefair P. J. (1991) Site layout planning for daylight and
sunlight. BRE Report BR209. Construction Research Communications, Garston, UK.
Littlefair P. J. (1993) Measuring daylight. Building Research
Establishment Information Paper IP 23 /93. Construction
Research Communications, Garston, UK.
Littlefair P. J. (1996) Designing with innovative daylighting.
BRE Report. Construction Research Communications, Garston, UK.
Littlefair P. J. and Aizlewood M. (1998) Daylighting in atrium
buildings. BRE Information Paper IP3 /98. Construction
Research Communications, Garston, UK.
Lynes J. A. (1979) A sequence for daylighting design. Light.
Res. Technol. 11(2), 102106.
Seshadri T. N. (1960) Equations of sky components with a
CIE standard overcast sky. Proc. Indian Acad. Sci. 57A,
233.
Ward G. J. (1994) The Radiance lighting simulation and
rendering system. Computer Graphics Proceedings Annual
Conference Series, pp. 4592472.