Professional Documents
Culture Documents
Rowan, Williams, Davies and Irwin (RWDI) Inc., 600 Southgate Dr., Guelph, ON, N1G 3W6, Canada
Rowan, Williams, Davies and Irwin (RWDI) Inc., 230 e 1385 West 8th Ave, Vancouver, BC, V5Z 1K5, Canada
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 23 February 2016
Received in revised form
28 March 2016
Accepted 12 April 2016
Available online 20 April 2016
In recent years there has been a growing awareness of the risks posed to people and property from
uncontrolled solar reections from the built environment. Despite the severity of the risks, there is
surprisingly little regulation regarding such reections. Presumably, part of the reason for this lack of
regulation is that there are no universally accepted criteria from the scientic community dening
acceptable limits of reected visible light and thermal irradiance in the urban realm. Without appropriate guidance, the regulations which are employed by cities may not be appropriate and designers have
no means to judge the impact of a potential building's reections until after its built. This paper presents
a review of existing regulations and metrics related to the impact of visible light and thermal energy on
people and property. It also proposes quantitative criteria which the authors have developed for use in
design and construction in order to help architects and designers understand the level of impact their
building's reections will have on its neighbors. The literature that the proposed criteria are based on is
still limited in breadth. It is our hope that the research and design communities will further develop the
criteria and tools that will benet designers and city regulators.
2016 Elsevier Ltd. All rights reserved.
Keywords:
Glare
Solar reection
Built environment
Regulation
Faade performance
1. Introduction
For people living in cities, seeing reected sunlight from a
building is a common experience, and while this can sometimes be
an annoyance, we mostly accept it as part of urban life. However,
buildings can cause more severe impacts owing to the type and
coverage of glass on the faade and also curvilinear designs. A wellpublicized example of signicant reections was the Disney Concert Hall in Los Angeles [1]. It had a signicant area of polished
stainless steel, which combined with the concave facade geometry,
would reect and sometimes focus light within the neighborhood.
This led to heat gain issues at sidewalks and inside surrounding
buildings. The bright reections also posed a visual distraction to
drivers. Ultimately, parts of the faade were roughened to help
scatter the reections, reducing their intensity. More recent examples of reected light causing distraction or damage include the
* Corresponding author.
E-mail addresses: ryan.danks@rwdi.com (R. Danks), joel.good@rwdi.com
(J. Good), ray.sinclair@rwdi.com (R. Sinclair).
http://dx.doi.org/10.1016/j.buildenv.2016.04.017
0360-1323/ 2016 Elsevier Ltd. All rights reserved.
Shard in London [2], Museum Tower in Dallas [3], the Vdara Hotel in
Las Vegas [4] and 20 Fenchurch in London [5]. The latter two
buildings reportedly focused enough sunlight in pedestrian areas to
cause burns on a guest and damage a car with radiant heat,
respectively. Buildings are not the only potential sources of
dangerous glare from the built environment; photovoltaic (PV)
panels [6] and even art installations [7] have been blamed for visual
and thermal issues related to reected sunlight as well.
The well-publicized nature of these events has led to increased
awareness in the design community, and a desire to understand
how a proposed building will interact with the sun. As consulting
engineers in the eld of building science, the authors have been
asked with increasing frequency: How will reections from my
building affect its neighbors? With little guidance available from
governing bodies we endeavored to create suitable criteria to assess
the visual and thermal impacts of reected light from the built
environment. These criteria were developed and rened through a
review of available literature across multiple scientic and engineering disciplines as well as through experience gained from the
analytical and physical study of reections from several buildings
around the world.
194
thermal energy to approximately the same degree. Today, highperformance glazing systems are available which often employ
low-emissivity coatings. These products can be signicantly more
reective to thermal energy than they are for visible light. As an
example, consider the common Solarban 70XL glazing system,
manufactured by PPG Industries. According to the manufacturer
[16], typical double-pane glazing systems using Solarban 70XL,
reect between 5% and 21% of exterior visible light, which mostly
fall within the prescribed reectance limits discussed earlier. But,
when the glazing's reectance is averaged over the full solar
spectrum (i.e. ultraviolet, visible and infrared radiation), the
reectance of some systems can increase dramatically. In the case
of Solarban 70XL employed with two panes of clear glass, the full
spectrum reectance is more than four times the visible reectance. This is an important nuance as the heating of an object will
be driven by both visible and thermal energy. The effect that
Solarban 70XL has on the visible and full spectrum reectance of a
single pane of 6 mm glass is also illustrated in Fig. 1.
Another shortfall of prescriptive regulations is that they typically do not consider the nature of the impacts (i.e., the who or
what being impacted), nor the frequency and duration of the
reections. A visible reection that falls on a pedestrian could be
benign in many situations but when it aligns with a vehicle driver's
required line of sight and impairs their vision, it becomes a danger.
Similarly, a short duration reection is unlikely to cause signicant
heating or damage to surrounding vegetation or man-made objects,
but the longer the exposure, the higher risk of the potential adverse
thermal impacts.
Finally, to stipulate a general upper limit of reectance to any
one unit of a glazing assembly does not account for the building's
shape. In particular, a concave shaped facade can potentially
concentrate reections in one area and a convex shape can cause
repeated reections over a longer period of the day over a broader
area of the surrounding neighborhood. As faade designs become
more complex, these impacts become harder to intuitively predict.
For example, a single surface adhering to a 20% maximum reectance limit may not cause problems, but if a number of these individual surfaces create a focal point (e.g. a faade with a concave
curvature, illustrated in Fig. 2 above) the thermal effects would be
additive and could ultimately cause thermal discomfort or burns on
people or damage to plants and materials. Situations such as these
are often the culprit in cases of extreme reection related damage,
which have been dubbed death rays by the media.
Due to these limitations with existing prescriptive limits, it is
the authors' opinion that a more qualitative, analytical approach be
taken to assessing the impact of urban building reections.
3. Review of existing reection impact metrics
A current limitation with regulating urban reections is that
there is a lack of widely accepted metrics for dening the full
impact of a building's reections on its surroundings. This is at least
partially a result of the subjectivity of how an individual experiences a bright light. This experience is dependent on a variety of
factors including what the person is doing, their expectations, their
age, etc. Therefore, there is currently no universally accepted metric
for the impact of day light even within the transportation and
lighting communities.
A separate issue from visual impacts is the impact of reected
thermal energy. There is limited research available on how much
reected energy is too much from a building. Beyond the issue of
intense focused reected thermal energy and its impact on human
safety and material damage, are more subtle questions about human comfort and even additional cooling loads for adjacent
buildings.
195
Fig. 2. Schematic of a curved faade in elevation (left) and plan (right) focusing and concentrating reected energy.
[18,19] have also shown that glare sources which would not be
considered sources of disability glare can still degrade driver performance. Additionally, many metrics related to disability glare are
intended for night lighting of roads, and therefore are not necessarily applicable in a day time setting. As an example, the
commonly used Threshold Increment (TI) metric, which can be
thought of as the increase in an object's contrast required to
maintain visibility after a glare source is introduced, is only valid for
average road luminance levels between 0.05 and 5.0 cd/m2 [20].
During the day, typical road luminance values are several orders of
magnitude above this range [21], making this metric invalid for use
with solar reections. Despite these challenges, several metrics
have been proposed based on disability glare research.
One of the more common metrics in industry comes from
Hassall [22], where the so called veiling luminance (a measure of
disability glare) is limited to no more than 500 cd/m2 for motorists.
There is an assumption in this limit that glare only occurs during
late dawn or early twilight time frames, when a driver's eyes would
196
One metric, the Daylight Glare Probability (DGP) [33] does aim
to alleviate this problem by including the total illuminance on the
eye in addition to the luminance ratio described above in its glare
impact prediction. This improves the ability of the metric to
correctly identify extremely bright lights (i.e. the sun) as glare
sources. However, like the other discomfort glare metrics, DGP has
been correlated to survey data for individuals adapted to indoor
ofce-like conditions. Studies would be required in order to assess
whether it would be appropriate for individuals located in outdoor
environments, with signicantly brighter ambient lighting conditions and for people conducting signicantly different activities.
Particularly as some research has shown that even an individual's
thermal comfort can inuence their glare perception [34].
3.1.3. Other visual glare metrics
Studies into the physiological response of individuals to glare
have shown that the reexive physical changes of the human eye
when exposed to bright light correlate well with perceived
discomfort [35,36]. This type of approach is advantageous because
there is lower variability in the geometry and behavior of healthy
adult eyes under daytime conditions [37e39] compared to the
variability in the adaptation luminance (i.e. the brightness level the
eye is adapted to). Adaptation luminance denes what is considered impactful glare in most other metrics and depends on task and
the preferences of an individual in addition to physiology, making it
challenging to model generally [40]. While one would still have to
correlate the physical response to the external stimuli, this area of
research has promise for application to exterior reections as it has
the potential to at least partially reduce the subjectivity of visual
glare impact classication by taking individual preference out of
consideration.
Ho et al. [41] have formulated a visual glare risk metric based on
the potential of a glare source to cause after images in an observer's eld-of-view or, in more extreme cases, retinal damage.
Aside from the size and intensity of the glare source, the primary
factors which dene this metric are the size of the pupil and the eye
as well as the speed of the human blink reex. As described above,
all these factors are relatively consistent for healthy, daylight
adapted adults. Most people have experienced the after-images
described by Ho when a camera ash is used in a dark room, which
also makes this metric very relatable for a lay-person compared to
some of the more subjective glare ranking scales [28]. This metric is
also easy to compute, allowing for rapid identication of problems
and mitigation tests. The use of this methodology is also explicitly
required by the Federal Aviation Administration as part of the
approval process for PV installations near airports [8] The main
limitations of the web-based tool also created by Ho [42] with
respect to its more general application, are that it does not identify
risks of multiple focused reections, or the specic location(s) of
the glare source(s). Also, the simplied geometrical analysis
employed in the tool ignores the impact of shading from the built
environment; which while appropriate for the open unobstructed
areas around airports, would be overly simplistic in an urban
context.
Attempts have also been made to leverage image processing
techniques to judge glare impacts on a viewer in terms of the
signal-to-noise ratio in a eld of view [43]. While this does have the
benet of being less subjective, further research is needed to dene
appropriate acceptability criteria.
3.2. Thermal impact of solar reections
In contrast to the plethora of metrics which assess the impacts
of visible light on individuals, there seems to be little if any publication of discussion of criteria for thermal energy reected from
buildings. To the authors' knowledge, the only available literature
regarding thermal irradiance limits is found in re safety or
military-related publications, referring to the impact on people and
property from the radiated energy of res or explosions.
3.2.1. Impacts on people
Fire safety research has provided a good understanding of how
long bare skin can be exposed to various levels of thermal radiation
before the onset of pain and then second degree burns. However,
there are many factors that can affect the severity and consequences of an exposure to elevated levels of thermal radiation.
Certainly, other environmental factors such as ambient temperature, clothing, age and health of person, and their ability to relocate
to less extreme exposure are just a few of the factors. Raj [44]
willingly exposed himself to 3500 W/m2 to 5000 W/m2 radiation
levels from a liqueed natural gas (LNG) re. He noted that with
exposures of several tens of seconds he did not suffer any severe
pain, burns or any other injury. He notes that American regulations
at that time had proposed an exposure limit of 1500 W/m2, for
which he argues are not justied due to attenuating effects of
typical clothing. The National Fire Protection Agency (NFPA) [45],
Society of Fire Protection Engineers (SFPE) [46] and British Standards Institution (BSI) [47] dene 2500 W/m2 as an upper limit of
acceptability for thermal radiation exposure during egress from a
re. They indicate that this level of intensity can be tolerated for
approximately 30 s before the onset of pain, allowing enough time
for an individual to evacuate. The Federal Emergency Management
Agency (FEMA) publishes a similar guideline for thermal irradiance
exposure [48].
The typical maximum level of solar irradiance individuals would
be exposed to naturally is approximately 1000 W/m2, and in the
authors' experience a typical intensity of a building's reection is in
the 200e400 W/m2 range, with peak intensities approaching
800 W/m2 possible. In contrast, Fig. 4 was produced during a study
of a building found to be focusing signicant amounts of solar
197
198
Fig. 4. Plots of measured and simulated irradiance levels on a horizontal surface for case study with focusing reections.
Fig. 5. Visible light and thermographic photographs of surfaces exposed to direct solar irradiance.
problems. The challenge in assessing the adverse impact of reections on plants is that they can be become unhealthy from a
wide variety of other factors, or a combination of factors, including:
excessive/insufcient pruning or mowing, pedestrian trafc, over/
under watering, poor quality soil, poor ventilation, and infestation
of insects, molds, or fungi. Moreover, the degree of the adverse
effects would vary by species of plant. This gap in current knowledge limits our ability to establish limiting criteria on reection
impacts on plants.
3.3. Additional metrics
In addition to metrics which intend to quantify the specic
impacts of visual and thermal reections, there are also other
proposed metrics which try to provide an indication of the reection impacts of a building in a more general manner. These types of
metrics are typically geometric in nature, using only the geometric
relationship between the sun and the reecting building to dene
the impact of reections. Shih and Huang [56] dene a boundary
of reection area (BRA) parameter, which is essentially the area of
a horizontal grade level plane which is exposed to reections from
the subject building. While simple to understand and compute, this
method fails to acknowledge the impact of variable glazing
reectivity, shading and the risk of focused reections. Brzezicki
suggests a similar analytic approach to provide an early warning
of potential reection focusing from curved facades [57]. As with
the BRA, while easy to compute, this approach ignores the variable
reectivity of common faade elements, shading and reections
from planar surfaces.
Another approach is to use forward ray tracing methods (i.e. a
computer simulation which tracks the paths of light rays starting at
the sun and moving forward in time) to identify areas within a
study domain where reections land, and then quantify the
reection impact in terms of the density of rays which impact a
given area [58], or in terms of the reected energy expressed as a
multiple of the incident light [59]. While the density of reected
rays impacting a surface does identify locations where focused,
frequent or long duration reections occur, it gives no indication of
the overall intensity of those reections. Similarly, proposing limits
on reected light to no more than ambient (as in Ref. [59]) can be
misleading. Ambient solar intensity varies over the course of a day
and throughout the year, a reection which is twice as intense as
the sun at dawn may be an order of magnitude less intense than a
reection which is twice as intense as the sun at noon and thus
have vastly different visual and thermal consequences for people
and property exposed to it.
4. Proposed criteria
The following sections dene a set of criteria that the authors
propose are suitable for dening the level of impact a reection will
have on its surroundings. The use of the proposed visual light and
thermal radiation criteria presented here presuppose that a reliable
method of measuring and/or predicting reected irradiance at a
given point has been used as an input. A review of the various
potential pieces of software which can be used to calculate reected
irradiance is a topic worthy of its own discussion, which the authors
plan to address in an upcoming publication [49].
The impacts of visual reections will occur at much lower intensities than those which can cause thermal impacts; also many
materials (particularly high performance architectural glazing) will
reect different amounts of visible and thermal irradiance. Therefore, we believe it is appropriate to dene separate criteria for the
visual and thermal impacts of reected solar energy to properly
understand both types of effects.
199
200
Table 1
Proposed visual impact categorization.
Visual impact
level
Reection conditions
None
Low
Moderate
High
Extreme
Table 2
Proposed thermal limits.
Condition
1500 W/m2
2500 W/m2
201
202
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]