Building and Environment 108 (2016) 110e121

Contents lists available at ScienceDirect

Building and Environment
journal homepage: www.elsevier.com/locate/buildenv

Effect of pavement thermal properties on mitigating urban heat
islands: A multi-scale modeling case study in Phoenix
Jiachuan Yang a, *, Zhi-Hua Wang a, Kamil E. Kaloush a, Heather Dylla b
a
b

School of Sustainable Engineering and the Built Environment, Arizona State University, Tempe, AZ 85287, USA
National Asphalt Pavement Association, Lanham, MD 20706, USA

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 16 June 2016
Received in revised form
15 August 2016
Accepted 20 August 2016
Available online 22 August 2016

Engineered pavements cover a large fraction of cities and offer signiﬁcant potential for urban heat island
mitigation. Though rapidly increasing research efforts have been devoted to the study of pavement
materials, thermal interactions between buildings and the ambient environment are mostly neglected. In
this study, numerical models featuring a realistic representation of building-environment thermal interactions, were applied to quantify the effect of pavements on the urban thermal environment at
multiple scales. It was found that performance of pavements inside the canyon was largely determined
by the canyon geometry. In a high-density residential area, modifying pavements had insigniﬁcant effect
on the wall temperature and building energy consumption. At a regional scale, various pavement types
were also found to have a limited cooling effect on land surface temperature and 2-m air temperature for
metropolitan Phoenix. In the context of global climate change, the effect of pavement was evaluated in
terms of the equivalent CO2 emission. Equivalent CO2 emission offset by reﬂective pavements in urban
canyons was only about 13.9e46.6% of that without building canopies, depending on the canyon geometry. This study revealed the importance of building-environment thermal interactions in determining thermal conditions inside the urban canopy.
© 2016 Elsevier Ltd. All rights reserved.

Keywords:
Building-environment thermal interactions
CO2 emission offset
Energy consumption
Outdoor thermal comfort
Pavement
Urban heat island mitigation

1. Introduction
Urban heat island (UHI), a presence of higher temperatures in an
urban area as compared to its rural surroundings, has been
considered as one of the major problems in the 21st century [1].
Using satellite-measured surface temperatures, a higher UHI intensity was usually found in summer than that in winter [2]. The
adverse effects induced by UHI include but are not limited to,
elevated temperatures [3], increased energy consumption [4], air
pollution [5], heat-related mortality [6], and disruption to ecosystems [7]. Under the challenge of future climate change, UHI makes
cities become unprecedentedly vulnerable to environmental
problems that research efforts have been devoted to developing
and testing adaptation/mitigation strategies during the past decades [8,9]. Recognized strategies include reﬂective roofs [10,11],
green roofs [12,13], urban vegetation and shading [14e16], heat
sinks [17], and cool pavements [18]. Due to the space restrictions in
the urban environment, roof material has been extensively studied

* Corresponding author.
E-mail address: jyang104@asu.edu (J. Yang).
http://dx.doi.org/10.1016/j.buildenv.2016.08.021
0360-1323/© 2016 Elsevier Ltd. All rights reserved.

while pavement material on the ground level has gained limited
attention.
Paved surfaces, including roads, parking areas and sidewalks,
cover a signiﬁcant percentage of urban surfaces. A previous study
reported that the percentage of paved surfaces ranged from 30 to
39% as seen from above the urban canopy, and from 36 to 45% as
seen from under the canopy for a variety of metropolitan areas [19].
With such a signiﬁcant percentage, modiﬁcation of pavement
materials provides a large potential for mitigating urban heat
islands. Though a couple of studies have illustrated the capacity of
pavements in reducing UHI and building energy consumption,
thermal interactions between buildings and the surrounding
microclimate in urban canopies are largely neglected [20]. Different
from roof materials, pavement are located inside the urban canyon
where buildings substantially alter the heat transport via shading
and reﬂecting radiations. With a three-dimensional building-tocanopy model, Yaghoobian and Kleissl [21] showed that the reﬂected solar radiation from the reﬂective pavement can increase
annual cooling loads of nearby ofﬁce buildings by up to 11% in
Phoenix. Li [22] conducted an experimental study in Davis, California and observed that around noon on sunny summer days, the

J. Yang et al. / Building and Environment 108 (2016) 110e121

temperature of building walls was 2e5  C hotter over concrete
pavements than over asphalt pavements. Incorporation of the
environmental complexity of built terrains is therefore of critical
importance in order to provide useful guidance to energy-efﬁcient
designs of buildings and paved surfaces, leading to sustainable city
planning. In addition, most existing studies have focused on retroﬁtting pavements with high albedo or reﬂective construction
materials, thus other pavement materials, such as ones with
different thermal capacity and conductivity, have been less studied
as strategies for UHI mitigation.
Existing studies on the impact of pavements on the urban
thermal environment focused on the building-resolving scales and
explored ofﬂine models where meteorological forcing is provided
as boundary conditions [20]. Geographical complexities at city and
regional scales, such as spatial heterogeneity, variability in building
geometry and density, and local air circulation, are not properly
represented. In addition, land-atmosphere interactions are largely
neglected in ofﬂine models, i.e., meteorological conditions do not
respond to changes in building physics. Due to theses limitations,
upscaling the results of ofﬂine studies for guidance at city and
regional scales becomes challenging [23]. Accurate quantiﬁcation of
the effect of pavements on the urban thermal environment
necessarily requires studies in a fully-interacting environment, i.e.,
a coupled atmosphere-urban modeling system.
In this study, the main objective was to evaluate the effect of
pavement thermal properties on the urban thermal environment at
multiple scales. Towards this end, numerical models were applied
with building-environment thermal interactions to: 1) analyze the
sensitivity of pavement surface temperature to canyon geometry, 2)
identify the impact of pavement thermal properties on the surface
temperature, building energy consumption, and outdoor human
thermal comfort at a neighborhood scale, and 3) assess the effect of
pavement thermal properties in mitigating UHI at a regional scale.
The metropolitan Phoenix in Arizona was selected as the testbed
because of the rapid urbanization in the past decades which has
created a signiﬁcant UHI in this region. The equivalent CO2 emission
offset by reﬂective pavement due to modiﬁed radiative forcing at
the global scale was also estimated and discussed.

111

canyon, and analytical solutions to heat transfer in building envelopes. Model performances over various pavement surfaces in the
urban area have been validated by in-situ measurements under
different climate conditions [25,26]. Detailed computational processes of the model can be found in the original paper and thus are
not duplicated here.
2.2. WRF-urban modeling system
During the past decades, a growing concern on urban heat island has led to development of numerous mesoscale atmosphereurban modeling systems [27e29]. Among the developed systems,
one powerful tool is the WRF-Urban modeling system, which has
been widely utilized and examined for major metropolitan regions
around the world [30e32]. In this study the developed urban
canopy model was implemented into the WRF model version 3.4.1
and the new model was employed to access the impact of pavements at the regional scale. Initial meteorological conditions for the
WRF simulations were obtained from the National Centers for
Environmental Prediction Final Operational Global Analysis data,
which were available on a 1  1 resolution with a 6-h temporal
frequency (details can be found on http://rda.ucar.edu/datasets/
ds083.2/). Land use information was acquired from the National
Land Cover Database (NLCD) 2006 [33].
2.3. Study area
For the urban canopy model, meteorological measurements in
the atmospheric boundary layer are required as model inputs.
Observations obtained from the eddy-covariance ﬂux tower
deployed at Maryvale, west Phoenix (see Fig. 1(b)) were used to
drive ofﬂine simulations. The experimental site is a high-density
residential area with single-family houses with a mean lot size of
about 700 m2 [34]. Generally, the daily mean incoming shortwave
radiation is greater than 600 W m2 and the daily mean air temperature at 22 m is higher than 30  C in summer. The monsoon
season starts June 29 and ends September 30. The wind speed in
the study area is about 3 m s1 in summer, with a prevailing wind
direction towards the west [34].

2. Numerical tools and study area
3. Effect of canyon geometry
This section describes the numerical tools used to investigate
impacts of pavements in the built environment in this study. An
ofﬂine (stand-alone) urban canopy model was adopted for studying
the impact at the neighborhood scale. The ofﬂine model is suitable
for long-term simulations at the neighborhood scale due to its high
computational efﬁciency. However, land-atmosphere interactions
are largely neglected in the ofﬂine model. At the city and regional
scales, an online (coupled) Weather Research and Forecasting
(WRF)-Urban modeling system was used to account for the landatmosphere interactions and surface heterogeneity. Nevertheless,
the requirement of high-performance computational resources
imposes constraints on spatial resolution and simulation time of
online models.
2.1. Urban canopy model
To accurately quantify the impact of pavements in the built
environment, a numerical model that captures coupled urban energy and water budgets is needed. Here a state-of-the-art urban
canopy model (UCM) [24,25] was used. The UCM represents
building arrays as a two-dimensional street canyon, as illustrated in
Fig. 1(a). The model features a realistic representation of hydrological processes over natural and engineered surfaces, sub-facet
heterogeneity, building-environment thermal interaction in the

Urban geometry plays a crucial role in determining the magnitude of shading and trapping effects that its impact on pavement
surface temperature needs to be addressed [35]. In this section the
sensitivity of pavement surface temperature to canyon geometry
was analyzed with the urban canopy model. Capability of the model
in reproducing energy and water budgets of the study residential
area at the annual scale (year 2012) was veriﬁed in a previous study
[36]. The calibrated parameters were summarized in Table 1 and
were used for subsequent ofﬂine numerical simulations. As
reﬂective pavement has gained increasing popularity recently, it
was used as an example to illustrate the effect of canyon geometry.
Focusing on the time when the urban thermal environment is
extremely aggravated, numerical simulations were carried out for a
pre-monsoon summer period, 12e17 June 2012, when high temperatures were observed under clear sky conditions. Pavements
were modeled with an albedo (the ratio of reﬂected radiation from
the surface to incident radiation upon it) of 0.1, 0.3, 0.5 and 0.7,
while other thermal properties remained the same as shown in
Table 1. The ﬁrst set of simulation modeled a pavement surface
without canopy (e.g., open ground space, parking lot), while the
second set simulated a pavement surface in the canyon of a typical
high-density residential area.
Averaged diurnal proﬁles of pavement surface temperature are

112

J. Yang et al. / Building and Environment 108 (2016) 110e121

Fig. 1. (a) Schematic of the urban canopy model: h, w, and r are the normalized building height, canyon width, and building width, respectively; H is the sensible heat ﬂux, G is the
storage heat ﬂux, T is the temperature, z is the height, subscripts R, W, G, can, a, B denote properties of roof, wall, ground, canyon, air and building. (b) Location of the ﬂux tower in a
high-density residential area in Phoenix (Google Earth basemap).

Table 1
Input parameters for a residential area in Phoenix using the urban canopy model.
Input parameters

Symbol

Values

Input parameters

Symbol

Values

Latitude (degree)
Longitude (degree)
Emissivity of ground ()
Emissivity of wall ()
Emissivity of roof ()
Albedo of ground ()
Albedo of wall ()
Albedo of roof ()
Thickness of wall (m)
Thickness of roof (m)
Canyon orientation

Lat
Lon

34.42
111.93
0.95
0.95
0.95
0.1
0.2
0.1
0.18
0.3
p/8

Normalized building height
Normalized canyon width
Heat conductivity of ground (W m1 K1)
Heat conductivity of wall (W m1 K1)
Heat conductivity of roof (W m1 K1)
Heat capacity of ground (MJ m3 K1)
Heat capacity of wall (MJ m3 K1)
Heat capacity of roof (MJ m3 K1)
Roughness length above roof (m)
Roughness length above ground (m)

h
w
kG
kW
kR
CG
CW
CR
Zm,R
Zm,G

0.5
0.5
0.9
1.3
1.2
2.1
1.5
1.9
0.01
0.05

3G
3W
3R

aG
aW
aR
dW
dR

qcan

shown in Fig. 2. Fig. 2 shows that increasing pavement albedo
signiﬁcantly reduced surface temperature in the absence of canopy.

With an increase of 0.6 in surface albedo, daily peak pavement
surface temperature decreased about 23  C in average in the

J. Yang et al. / Building and Environment 108 (2016) 110e121

113

during the day was dissipated slowly.
A third set of simulation was carried out to quantify the effect of
canyon aspect ratio (h/w) and canyon orientation (qcan) on pavement surface temperature. In the simulation the albedo of pavement was increased from 0.1 to 0.7 with varying canyon aspect
ratios and orientations, while other parameters remained the same
as shown in Table 1. Fig. 3(a) presents the averaged diurnal proﬁles
of surface temperature reduction by increasing pavement albedo
from 0.1 to 0.7 with different canyon aspect ratios (qcan ¼ p/8). It
suggested that cooling effect of reﬂective pavement decreased with
aspect ratio. Daily peak temperature reduction by reﬂective pavement was about 27  C with an aspect ratio of 0.5, and became
around 14  C when h/w equaled to 4. In a deeper street canyon
(large h/w value), shading effect of buildings dominates that
pavement receives incoming solar radiation only around noon time.
Under this condition, the effective period of the reﬂective pavement is signiﬁcantly reduced. For example, with an aspect ratio of 4,
the cooling effect by the reﬂective pavement was larger than 4  C
only for the period from 1200 to 1530 local time. The impact of
canyon orientations is shown in Fig. 3(b) (h/w ¼ 1). Changing the
canyon orientation qcan modiﬁed the pattern of diurnal variation of
the cooling effect. For the study area in Phoenix, diurnal proﬁle of
the cooling effect with qcan ¼ p/8 was almost identical to that with

Fig. 2. Averaged diurnal proﬁles of predicted pavement surface temperatures with
different albedo (a) without a canopy, and (b) in a high-density residential area in
Phoenix during 12e17 June 2012.

simulation period. With the canopy, the reﬂective material was also
able to reduce pavement surface temperature considerably, with a
daily peak cooling effect of about 20  C. However, adjacent buildings blocked part of the incoming solar radiation for the pavement
and therefore the albedo was less effective inside the street canyon.
Without canopy, temperature difference between pavements with
different albedo increased rapidly after sunrise at around 0530 local
time. Increasing albedo from 0.1 to 0.7 reduced pavement surface
temperature by more than 10  C for about 11 h. On the other hand,
the albedo modiﬁcation did not cool pavement surface signiﬁcantly
until about 0930 local time with the canopy, when pavement
started to become not fully shaded by adjacent buildings. Cooling
effect of more than 10  C only lasted for about 5 h, which was less
than half of that without canopy. During nighttime, albedo of materials became ineffective due to the absence of solar radiation.
However, as more radiation was reﬂected and less thermal energy
was stored during the day, reﬂective pavement led to a reduction of
about 3  C in nighttime surface temperature both with and without
canopy. It is noteworthy that the nighttime surface temperature of
pavement was higher with the canopy due to radiative trapping.
Buildings absorbed outgoing longwave radiation from pavement
and emitted longwave radiation back. Thermal energy was thus
trapped inside the urban canyon that heat storage in the pavement

Fig. 3. Averaged diurnal proﬁles of surface temperature reductions by increasing the
pavement albedo from 0.1 to 0.7 with different (a) aspect ratios (qcan ¼ p/8), and (b)
canyon orientations (h/w ¼ 1) in Phoenix during 12e17 June 2012.

114

J. Yang et al. / Building and Environment 108 (2016) 110e121

qcan ¼ 7p/8, except that the magnitude was consistently smaller.

The efﬁcient period of the reﬂective pavement (cooling > 5  C) with
qcan ¼ 5p/8 was substantially longer than that with qcan ¼ p/8,
though the peak cooling effect was close. Among the studied
canyon orientations, the maximum daily peak temperature
reduction was observed with qcan ¼ 3p/8, which was about 7  C
greater than that with qcan ¼ 7p/8. Results indicated that when
deployed inside a street canyon, the performance of reﬂective
pavements is largely determined by canyon geometry.
4. Impact of pavements at a neighborhood scale
4.1. Surface temperature
To identify the importance of building-environment thermal
interactions in the urban canyon, in this section the effect of
pavements on the road and wall surface temperatures in Phoenix
was investigated. The sunny pre-monsoon period studied in section
3, 12e17 June 2012, was used for this section. To illustrate the
maximum effect of pavements, it was assumed that the road surface was 100% pavement with no trees or vegetation. For subsequent analysis in this section, an aspect ratio of 1 and a canyon
orientation of 0 were used for the study neighborhood.

Fig. 4(a) illustrates that increasing pavement albedo cooled the
road surface considerably. An increase of 0.6 in albedo led to a
pavement surface temperature reduction of up to about 20  C in
daytime. With reduced road surface temperatures, the reﬂective
pavement emitted less longwave radiation to the canyon environment. Nevertheless, more shortwave radiation was reﬂected that
the combined effect of the reﬂective pavement was complicated.
Fig. 4(b) shows that across the diurnal cycle, increasing pavement
albedo resulted in a warming effect on building walls from around
1030 to 1500 local time. The increased wall temperature indicated
that reﬂected shortwave radiation outweighed reduced longwave
radiation from the reﬂective pavement during this period. In the
rest of the diurnal cycle, the reﬂective pavement led to cooling of
wall surface temperatures. The maximum warming effect of about
2.0  C was observed at 1330 local time, while the maximum cooling
effect of about 1.9  C occurred at 2100 local time. Note that the
daytime warming effect may increase the peak electricity load and
cooling demand of the building, which is unfavorable for cities
especially in summers. In terms of nighttime cooling, it saves
cooling demands in hot seasons. However, it may cause additional
heating demands in cold seasons.
Performance of pavements depends on a number of thermal
properties, among those albedo is an important but not the only
determinant one. Pavements with high porosity, heat capacity, and
thermal conductivity have also been studied as strategies for UHI
mitigation [18]. Pavement with high thermal conductivity, and
pavement with large heat capacity were studied here as alternative
strategies for Phoenix urban heat island. Three simulations were
conducted with different values assigned to both thermal properties. For thermal conductivity, the studied values were 1.2, 2.0, and
3.0 W K1 m1, respectively. Tested heat capacities were 1.4, 2.0,
and 2.8 MJ K1 m3.
Results of thermal conductivity are plotted in Fig. 5. Fig. 5(a)
illustrates that increasing thermal conductivity led to lower road
surface temperatures during daytime for Phoenix, as a large thermal conductivity favored the transfer of radiative energy into the
pavement. On the other hand, during the nocturnal cycle, heat
released from the pavement due to a large thermal conductivity
increased the road surface temperature. This indicated that the
diurnal variation of road surface temperatures increased with the
rate of heat transfer process in the pavement. With an increase of
1.8 W K1 m1 in the thermal conductivity, maximum daytime
reduction and nighttime increment of the road surface temperature
were about 7.0 and 2.4  C, respectively. In terms of the wall surface
temperature, modifying the thermal conductivity of ground pavement had a relatively insigniﬁcant impact. Maximum cooling and
warming effects were about 1.0 and 1.3  C, respectively.
The effect of heat capacity is shown in Fig. 6, which was similar
to the result of thermal conductivity. Enabling ground pavements
to store more energy, a large heat capacity tended to reduce the
daytime peak road surface temperature. The stored energy was
released at night and led to a heating of the road surface. When the
heat capacity of ground pavements was doubled from 1.4 to
2.8 MJ K1 m3, maximum daytime reduction and nighttime
increment of the road surface temperature were about 5.7 and
2.2  C for Phoenix. With respect to the wall surface temperature,
modifying the heat capacity of ground pavement led to insigniﬁcant cooling and warming effects of up to 1.0 and 0.9  C. Simulation
results in this section illustrated that modifying the thermal
properties of pavements had a limited impact on the wall surface
temperature, even during the hot summer of a year in a desert city.
4.2. Building energy consumption

Fig. 4. Simulated (a) road surface temperatures, and (b) wall surface temperatures
with different surface albedo for the pavement in Phoenix during 12e17 June 2012.

After realizing the effect of different pavement thermal

J. Yang et al. / Building and Environment 108 (2016) 110e121

Fig. 5. Simulated (a) road surface temperatures, and (b) wall surface temperatures
with different thermal conductivity for the pavement in Phoenix during 12e17 June
2012.

115

Fig. 6. Simulated (a) road surface temperatures, and (b) wall surface temperatures
with different heat capacity for the pavement in Phoenix during 12e17 June 2012.

by:
properties on surface temperatures in the urban canyon, a series of
simulation was carried out to access the impact of pavement
thermal properties on the building energy consumption. Using the
calibrated UCM, the effect of pavement thermal properties on the
building energy consumption was tested at the annual scale. To
illustrate the effect of same materials on different urban facets,
simulations were conducted separately for roof and ground surfaces. Insulation was not considered in this study, which may lead
to an overestimation of the building energy consumption. For
ground and roof surfaces, four cases were studied. Thermal properties of the material in each case were (1) control case: a ¼ 0.1,
k ¼ 1.2 W K1 m1, C ¼ 1.4 MJ K1 m3; (2) reﬂective material:
a ¼ 0.5, k ¼ 1.2 W K1 m1, C ¼ 1.4 MJ K1 m3; (3) material of low
thermal conductivity: a ¼ 0.1, k ¼ 0.6 W K1 m1,
C ¼ 1.4 MJ K1 m3; and (4) material of large heat capacity: a ¼ 0.1,
k ¼ 1.2 W K1 m1, C ¼ 2.8 MJ K1 m3. These values were within
the typical range of engineering materials documented in the
literature. The thickness of materials for roof, wall, and ground
were 0.3, 0.18, and 0.45 m, respectively.
Following previous studies [36,37], the building energy consumption was estimated based on the heat ﬂux entering the
building through the innermost layer of the wall and the roof, given

Qin ¼

kðTin  TB Þ
;
din =2

(1)

where Qin is the heat ﬂux entering the building via wall or roof, din
and Tin are the thickness and temperature of the innermost
(discrete) layer, TB is the building indoor temperature. With this
equation, the estimated energy consumption mainly considers the
sensible load of buildings that the latent load is largely ignored. In
addition, two factors are neglected: (1) the internal energy loads
caused by people and equipment, and (2) the efﬁciency of air
conditioning system and the variation of the building indoor temperature. Note that heat ﬂuxes into the building through roof and
wall via conduction were calculated separately in the UCM. As U.S.
Department of Energy suggested that the thermostat should be set
to around 25.6  C for indoor thermal comfort [38], the building
indoor temperature was assumed to be maintained at 25  C by
indoor heating, ventilation, and air-conditioning (HVAC) systems
for the entire simulation period.
Fig. 7 presents the monthly energy consumption of the building
in a residential area in Phoenix. Total annual energy consumptions
due to the heat conduction via roof and wall with different

116

J. Yang et al. / Building and Environment 108 (2016) 110e121

Fig. 7. Monthly energy consumption of buildings due to heat conduction via (a) roof, and (b) wall with different engineering materials for Phoenix in 2012.

materials were summarized in Table 2. Note that in the UCM, the
rooftop is not thermally interacted with the street canyon. Therefore Fig. 7(a) represents the effect of various roof materials on the
building energy consumption, and Fig. 7(b) illustrates the impact of
pavement materials. In terms of the roof surface, Fig. 7(a) indicates
that the material of low thermal conductivity was the most efﬁcient
in reducing building energy consumptions. As it slowed the heat
exchange across the building envelope, energy consumptions were
substantially reduced throughout the year. Total annual energy
consumptions decreased by about 50% as compared to the control
case. Increasing albedo and heat capacity of the roof material also
led to evident savings of energy consumptions. In cool to cold periods, the reﬂective material yielded greater energy consumptions
than those with a large heat capacity, primarily due to the resulted
heating penalty at nighttime. Whereas in summers, warming effects from the material of large heat capacity caused extra cooling
demands at night, and the reﬂective material was more effective in
saving building energy consumptions. At the annual scale,
increasing the heat capacity of the roof material was more efﬁcient
in enhancing the building energy efﬁciency than increasing albedo
for the study area. Compared to deployment of reﬂective materials,
using materials of large heat capacity for the roof surface saved
total annual energy consumptions by about 9 kWh m2. This result
emphasizes the previous ﬁnding that the thermal performance of

materials is determined by multiple parameters, beneﬁts of various
mitigation strategies should be compared to come up with the best
solution [39].
Fig. 7(b) shows that energy consumptions due to the heat conduction via wall was signiﬁcantly larger than that via roof, mainly
due to the difference in the thickness of materials; however, among
studied thermal properties, it was found that modifying ground
pavements had negligible effects on the building energy consumption. Note that the unit of energy consumptions is kWh per
square meter of wall surface facing the sun. Total annual energy
savings in different cases were very small, ranging from 20 to
7 kWh m2, which was less than 3% of the annual energy consumption in the control case. Note that the material of low thermal
conductivity on roof substantially reduced building energy consumptions, however, it slightly increased the energy consumption
of adjacent buildings when deployed at the ground level. The results illustrate that the same engineering material may lead to an
opposite effect on the building energy efﬁciency when applied to
different urban spatial locations.
4.3. Outdoor thermal comfort
Under the semi-arid climate, residents experience intense
thermal discomfort during hot days in outdoor or non-air-

Table 2
Summary of thermal properties and total annual energy consumptions in different simulation cases for Phoenix in 2012.

Albedo
Thermal conductivity (W K1 m1)
Heat capacity (MJ K1 m3)
Total annual energy consumption through roof (kWh m2)
Total annual energy consumption through wall (kWh m2)

Control case

Reﬂective material

Material of low thermal conductivity

Material of large heat capacity

0.1
1.2
1.4
339
727

0.5
1.2
1.4
238
707

0.1
0.6
1.4
168
734

0.1
1.2
2.8
228
719

J. Yang et al. / Building and Environment 108 (2016) 110e121

conditioned indoor environments in metropolitan Phoenix [40].
High temperature signiﬁcantly increases the risk of heat-related
morbidity and mortality in cities, especially under the challenge
of global warming [41]. Close to pedestrians on the street level,
pavements play a dominant role on outdoor human thermal comfort. Quantifying thermal comfort in outdoor urban areas can be a
challenging process, due to the variety of environmental factors,
including temperature, humidity, wind speed, radiative exposure,
ambient evaporative, and sensible ﬂuxes [42]. The Index of Thermal
Stress (ITS) developed by Givoni [43] was chosen to assess the
impact of pavements on outdoor thermal comfort for Phoenix:

ITS ¼ ½Rn þ C þ ðM  WÞ=f ;

(2)

where Rn is the body's net radiation, C is the convective energy
exchange, M is the heat generated by body's metabolism, W is the
heat consumed through work, and f is the efﬁciency of sweat
evaporation. All variables in Equation (2) has the unit of W m2
except the dimensionless variable f. The ITS is a measure of the rate
at which the human body must secrete sweat to maintain thermal
equilibrium. It accounts for human metabolism, mechanical work,
radiation exchange, clothing condition, wind speed, humidity, and
temperature. By deﬁnition, outdoor thermal comfort decreases
with the ITS.
In this study, we assumed a pedestrian was performing mild
outdoor activities (e.g., walking) with light summer clothing in the
street canyon. In total four cases were studied for the simulation
period of 12e17 June 2012. Thermal properties of the pavement
surface in each case were the same as those listed in Table 2. Increase of the ITS by different pavements as compared to the control
case for the entire simulation period is presented in Fig. 8. Fig. 8
clearly demonstrates that modifying the thermal conductivity and
heat capacity of pavements had negligible impacts on outdoor
human thermal comfort across the diurnal cycle. On the other hand,
with increased solar reﬂection, the reﬂective pavement signiﬁcantly increased the thermal stress on pedestrians, up to more than
100 W m2 around noontime throughout the simulation period.
This ﬁnding was in agreement with recent studies where the
reduced surface temperature was found not enough to offset
increased radiation loads from reﬂective pavements [44]. In
nighttime, the effect of reﬂective pavements was insigniﬁcant as
solar radiation was not available.
4.4. Regional effect of pavements
To incorporate the impact of land-atmosphere interaction and
land surface heterogeneity, we adopted the coupled WRF-UCM
modeling system to quantify the regional effect of pavements for

117

metropolitan Phoenix. Following a previous study's setup [13], we
used an inner domain of 212 km  212 km to cover metropolitan
Phoenix with a spatial resolution of 2 km. Detailed experimental
set-up of the WRF-UCM modeling system is referred to the previous study [13]. Due to the limitation of time and computational
resources, the simulation period was set to be 1e11 July 2006. Using
the 2006 NLCD land cover data, the urban land-use was divided
into three categories: low-density residential, high-density residential, and industrial and commercial. The performance of the
WRF model against ﬁeld observations from ground-based weather
stations for Phoenix has been veriﬁed [13]. The calibrated parameters for each land-use category were summarized in Table 3.
Thermal properties of pavements in four studied cases were the
same as those listed in Table 2 except that the pavement albedo was
0.17 in the control case. Pavement properties were changed for all
three urban categories in the simulation. Simulated impact of the
reﬂective pavements on land surface temperature (Ts) and 2-m air
temperature (T2) at 1400 local time is shown in Fig. 9. As shown in
Fig. 9, increasing the albedo of pavement surfaces from 0.17 to 0.5
led to a cooling effect of less than 0.5  C for Phoenix. This temperature reduction was insigniﬁcant given that the surface temperature was higher than 45  C at 1400 local time in Phoenix, and
was signiﬁcantly smaller than the predicted reduction from ofﬂine
simulations. With limited cooling impacts on urban land surface,
the reﬂective pavement had negligible inﬂuences on 2-m air
temperature.
Simulated impact of the pavements with large heat capacity and
low thermal conductivity on land surface temperature for Phoenix
are plotted in Fig. 10. Increasing the heat capacity resulted in a
lower pavement surface temperature, and the pavement of reduced
thermal conductivity led to an elevated surface temperature. The
magnitude of temperature change in both cases was less than
0.5  C, indicating a limited effect of the pavements on land surface
temperature. In terms of 2-m air temperature, the effect of both
pavements was also negligible (results not shown here).
Online simulation results were qualitatively similar to ofﬂine
ﬁndings. However, the magnitude of the temperature change from
online simulations was substantially smaller than that predicted by
the ofﬂine model. The principal reason was that land surface
temperature from the WRF-UCM modeling system was a lumped
average over all urban surfaces including road, wall, and roof. Existence of bare soil and vegetative surface in urban areas further
attenuated the impact of pavements, which could be up to 35% in
low-density residential areas as shown in Table 3. After accounting
for the land-atmosphere interaction and land surface heterogeneity, modifying the thermal properties of pavements was found to
have a limited effect on urban land surface temperature and air
temperature for Phoenix at the regional scale.

Fig. 8. Increase in ITS of a pedestrian in the street canyon by different pavements during 12e17 June 2012 for Phoenix.

118

J. Yang et al. / Building and Environment 108 (2016) 110e121

Table 3
Summary of input parameters for different urban-land categories in the WRF-Urban modeling system for metropolitan Phoenix.
Input parameters

Unit

Industrial and commercial

High-density residential

Low-density residential

h (building height)
lroof (roof width)
lroad (road width)
furb (urban fraction)
aR (albedo of roof)
aW (albedo of wall)
aG (albedo of ground)
CR (heat capacity of roof)
CW (heat capacity of wall)
CG (heat capacity of ground)
kR (thermal conductivity of roof)
kW (thermal conductivity of wall)
kG (thermal conductivity of ground)

m
m
m
e
e
e
e
MJ K1 m3
MJ K1 m3
MJ K1 m3
W K1 m1
W K1 m1
W K1 m1

17.0
10.0
10.0
0.95
0.15
0.15
0.17
1.35
1.35
1.4
0.70
0.70
1.2

7.5
9.4
9.4
0.85
0.16
0.16
0.17
1.35
1.35
1.4
0.70
0.70
1.2

5.0
8.3
8.3
0.65
0.16
0.16
0.17
1.35
1.35
1.4
0.70
0.70
1.2

Fig. 9. Simulated impact of the reﬂective pavement on (a) land surface temperature,
and (b) 2-m air temperature at 1400 local time for Phoenix.

Fig. 10. Simulated impact of (a) the pavement of large heat capacity and (b) the
pavement of low thermal conductivity on land surface temperature at 1400 local time
for Phoenix.

5. CO2 emission offset
At the global scale, the ultimate goal of comparing effects of
different pavements is to determine the efﬁcient material for
countering global warming. A previous study [45] reported that a
net albedo increase of about 0.1 in global urban areas could induce a
negative radiative forcing on the earth equivalent to a one-time
offsetting 44 Gt of CO2 emissions. 40% of the emission offset

(20 Gt) was achieved by modifying pavement albedo, and the rest
60% (24 Gt) was accomplished through reﬂective roofs. Nevertheless, the neglect of thermal interactions inside the urban canyon
raises uncertainty for their ﬁndings. The UCM used in this study
also enables a more realistic estimate of the effect of pavements in
terms of CO2 emission offset. As heat capacity and thermal conductivity have no impact on shortwave radiation budgets, only

J. Yang et al. / Building and Environment 108 (2016) 110e121

reﬂective pavements were studied in this section.
The net shortwave radiation budgets over road and wall surfaces
inside the urban canyon are given by:


l
w  lshadow
aG FWG
SW ¼ ð1  aW Þ shadow þ
2h
w

l
þ shadow aW FWW SY ;
2h


w  lshadow lshadow
þ
aW FWW SY ;
SG ¼ ð1  aG Þ
w
2h

(3)

(4)

where S is the net shortwave radiation and SY is the download
incoming shortwave radiation; subscripts ‘W’ and ‘G’ denote wall
and ground respectively; lshadow is the normalized shadow length;
FWG and FWW are the sky view factors between wall and ground, and
between wall and wall, respectively, calculated as:

FWW

rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ
w2 w
¼ 1þ
 ;
h
h

FWG ¼ 0:5ð1  FWW Þ:

(5)
(6)

Equations (2)e(5) suggest that the magnitude of outgoing
shortwave radiation varies with urban geometry, albedo of ground
and wall surfaces, and normalized shadow length. Thus the

Fig. 11. Diurnal proﬁles of incoming and reﬂected shortwave radiation for pavements
with different albedo in Phoenix urban canyon on the summer solstice day, 2012.

119

sensitivity of reﬂected radiation (from urban canopy to overlying
atmosphere) to urban geometry and pavement albedo was investigated for the Phoenix metropolitan area in this section. Other
parameters were obtained from Table 1.
Note that the normalized shadow length varies with the solar
azimuth angle and zenith angle, therefore the resultant effect of
reﬂective pavements has diurnal and seasonal variations. Following
the previous study [45], a 0.15 increase in pavement albedo was
assumed for the Phoenix metropolitan area. Diurnal proﬁles of
incoming and reﬂected shortwave radiations for pavements with
different canyon aspect ratios are shown in Fig. 11 for the summer
solstice day, where a large variability was observed. After sunrise,
the albedo of pavement surface began to function by reﬂecting the
reﬂected shortwave radiation from building walls. As the sun
continued to rise, pavements began to receive both direct incoming
shortwave radiation from the sun and reﬂected shortwave radiation from walls. During this period pavement albedo became efﬁcient because direct incoming shortwave radiation was
signiﬁcantly larger than reﬂected shortwave radiation from walls.
Around noontime when the sun was directly overhead, pavements
received only direct incoming shortwave radiation from the sun
that the impact of pavement albedo inside the urban canopy was
similar to that of an open space without canopy in terms of radiative heat exchange. Despite variations of the canyon aspect ratio, a
0.15 increase in pavement albedo increased the reﬂected shortwave
radiation by about 180 W m2 around noontime for all cases.
Depending on the canyon aspect ratio, the efﬁcient period of
reﬂective pavement ranged from 3 h to 7 h.
To identify the seasonal variation of the impact of reﬂective
pavements on urban shortwave budgets, we employed an index
named daily effective albedo. Daily effective albedo is deﬁned as the
ratio between total reﬂected shortwave radiation and total
incoming shortwave radiation over an urban canyon across the
diurnal cycle. Compared to the material (nominal) albedo, the daily
effective albedo is more appropriate for accessing the effect of
pavements on urban radiation budgets as it accounts for thermal
interaction and geometry effect in cities. Variations of the daily
effective albedo for pavements in urban canyons with different
aspect ratios is plotted in Fig. 12 for year 2012. Fig. 12 shows that the
daily effective albedo had a larger seasonal variation in a deeper
canyon. Averaging over the entire year 2012, an increase of 0.15 in
pavement albedo resulted in an increase of 0.070, 0.045, and 0.021
in the daily effective albedo when h/w equaled to 1, 2, and 4,
respectively. These values were used for calculating the effect of
reﬂective pavements on CO2 emission offset at the global scale.
Following the methodology used by Akbari et al. [45], the
change in radiative forcing is 1.27 W m2 per 0.01 increase in the
albedo of global land surface. Considering a global paved surface

Fig. 12. Seasonal variation of the daily effective albedo for pavements with material albedo of 0.17 and 0.32 in Phoenix urban canyon in year 2012.

120

J. Yang et al. / Building and Environment 108 (2016) 110e121

area of 5.3  1011 m2 [45], the increase of 0.021e0.070 in average
daily effective albedo of urban areas is equivalent to an increase of
4.2  105 e 1.4  104 in the Earth's albedo. This corresponds to a
radiative forcing change of 5.33e17.78 MW and a one-time global
CO2 offset potential of about 2.78e9.33 Gt, which is about
13.9e46.6% of the 20 Gt estimated without considering thermal
interactions in the urban canopy. The presented results highlight
that the impact of reﬂective pavements largely depends on the
canyon geometry. Moreover, estimating its potential for offsetting
greenhouse emission at regional and global scales requires special
attention to the heterogeneity of canyon geometry in different
cities.
6. Conclusion
This study adopted numerical models with buildingenvironment thermal interactions to evaluate the effect of pavement thermal properties on urban environment at multiple scales.
Various thermal properties of pavements were tested, including (1)
reﬂectivity, (2) heat capacity, and (3) thermal conductivity. Results
showed that the effect of pavement thermal properties on road and
wall surface temperatures was largely modulated by the canyon
geometry. Due to radiative shading effect, modifying thermal
properties of pavements had a relatively insigniﬁcant impact in the
urban canyon as compared to that without canopy. For a highdensity residential area in Phoenix, different pavements led to
similar wall temperatures and building energy consumptions.
When applied to different urban facets, same engineering material
may have different effects on the building energy efﬁciency. In
terms of outdoor human thermal comfort, due to increased reﬂected radiation, reﬂective pavements increased the thermal stress
of pedestrians inside the urban canyon considerably, while other
pavements did not cause this adverse effect.
With dynamic atmospheric modules and detailed land use land
cover information, the online simulation using the WRF-urban
modeling system illustrated that modifying pavements had
limited impacts on urban land surface temperature and air temperature for the Phoenix metropolitan area. Considering buildingenvironment thermal interactions, a more realistic impact of
reﬂective pavements in terms of CO2 emission offset and radiative
forcing change was estimated. Increase in the daily effective albedo
was found to be signiﬁcantly smaller than the modiﬁcation of
pavement albedo. Depending on the canyon aspect ratio, modifying
pavement albedo in the urban canyon led to an offset 13.9e46.6% of
the previous estimate, which did not account for buildingenvironment thermal interactions.
Results of numerical simulations suggest that buildingenvironment thermal interactions play a crucial role in determining thermal conditions in urban areas. Accounting thermal interactions in a built environment leads to more realistic and reliable
results of numerical simulations. This study provides information
regarding how reﬂective pavements are not an exceptionally
effective strategy for mitigating urban heat island and enhancing
building energy efﬁciency as suggested by previous studies that did
not consider building-environment thermal interactions.
Acknowledgement
This work is supported by the National Asphalt Pavement Association and US National Science Foundation under grant number
CBET-1435881.
References
[1] A.M. Rizwan, L.Y.C. Dennis, C. Liu, A review on the generation, determination

and mitigation of Urban Heat Island, J. Environ. Sci. 20 (1) (2008) 120e128.
[2] H. Tran, D. Uchihama, S. Ochi, Y. Yasuoka, Assessment with satellite data of the
urban heat island effects in Asian mega cities, Int. J. Appl. Earth Obs. Geoinform. 8 (1) (2006) 34e48.
[3] C. Wang, S.W. Myint, Z. Wang, J. Song, Spatio-temporal modeling of the urban
heat island in the Phoenix metropolitan area: land use change implications,
Remote Sens. 8 (3) (2016) 185.
[4] H. Akbari, Cooling Our Communities. A Guidebook on Tree Planting and Lightcolored Surfacing, Lawrence Berkeley National Laboratory, 2009.
[5] C. Sarrat, A. Lemonsu, V. Masson, D. Guedalia, Impact of urban heat island on
regional atmospheric pollution, Atmos. Environ. 40 (10) (2006) 1743e1758.
[6] J. Tan, Y. Zheng, X. Tang, C. Guo, L. Li, G. Song, et al., The urban heat island and
its impact on heat waves and human health in Shanghai, Int. J. Biometeorol. 54
(1) (2010) 75e84.
[7] P. Bolund, S. Hunhammar, Ecosystem services in urban areas, Ecol. Econ. 29 (2)
(1999) 293e301.
[8] Wong E, Akbari H, Bell R, Cole D. Reducing urban heat islands: compendium of
strategies. Environ. Prot. Agency, Retrieved June 2016.
[9] A.J. Arnﬁeld, Two decades of urban climate research: a review of turbulence,
exchanges of energy and water, and the urban heat island, Int. J. Climatol. 23
(1) (2003) 1e26.
[10] H. Akbari, R. Levinson, Evolution of cool-roof standards in the US, Adv. Build.
Energy Res. 2 (1) (2008) 1e32.
[11] M. Santamouris, Cooling the citiesea review of reﬂective and green roof
mitigation technologies to ﬁght heat island and improve comfort in urban
environments, Sol. Energy 103 (2014) 682e703.
[12] J. Yang, Z.-H. Wang, Physical parameterization and sensitivity of urban hydrological models: application to green roof systems, Build. Environ. 75
(2014) 250e263.
[13] J. Yang, Z.-H. Wang, M. Georgescu, F. Chen, M. Tewari, Assessing the impact of
enhanced hydrological processes on urban hydrometeorology with application to two cities in contrasting climates, J. Hydrometeorol. 17 (4) (2016)
1031e1047.
[14] H. Akbari, M. Pomerantz, H. Taha, Cool surfaces and shade trees to reduce
energy use and improve air quality in urban areas, Sol. Energy 70 (3) (2001)
295e310.
[15] L. Shashua-Bar, M. Hoffman, Vegetation as a climatic component in the design
of an urban street: an empirical model for predicting the cooling effect of
urban green areas with trees, Energy Build. 31 (3) (2000) 221e235.
[16] Z.-H. Wang, X. Zhao, J. Yang, J. Song, Cooling and energy saving potentials of
shade trees and urban lawns in a desert city, Appl. Energy 161 (2016)
437e444.
[17] G. Mihalakakou, M. Santamouris, D. Asimakopoulos, Use of the ground for
heat dissipation, Energy 19 (1) (1994) 17e25.
[18] M. Santamouris, Using cool pavements as a mitigation strategy to ﬁght urban
heat islandda review of the actual developments, Renew. Sustain. Energy
Rev. 26 (2013) 224e240.
[19] K.A. Gray, M.E. Finster, The Urban Heat Island, Photochemical Smog, and
Chicago: Local Features of the Problem and Solution, Northwestern University, Department of Civil Engineering, 1999.
[20] J. Yang, Z. Wang, K.E. Kaloush, Unintended Consequences: a Research Synthesis Examining the Use of Reﬂective Pavements to Mitigate the Urban Heat
Island Effect, 2013.
[21] N. Yaghoobian, J. Kleissl, Effect of reﬂective pavements on building energy use,
Urban Clim. 2 (2012) 25e42.
[22] H. Li, Evaluation of Cool Pavement Strategies for Heat Island Mitigation [Ph.D.
Thesis], Davis: University of California, 2013, p. 367.
[23] P. Ramamurthy, E. Bou-Zeid, J.A. Smith, Z. Wang, M.L. Baeck, N.Z. Saliendra, et
al., Inﬂuence of subfacet heterogeneity and material properties on the urban
surface energy budget, J. Appl. Meteorol. Climatol. 53 (9) (2014) 2114e2129.
[24] Z.H. Wang, E. Bou-Zeid, J.A. Smith, A coupled energy transport and hydrological model for urban canopies evaluated using a wireless sensor network,
Q. J. R. Meteorol. Soc. 139 (675) (2013) 1643e1657.
[25] Z.-H. Wang, E. Bou-Zeid, J.A. Smith, A spatially-analytical scheme for surface
temperatures and conductive heat ﬂuxes in urban canopy models, Bound.
Layer Meteorol. 138 (2) (2011) 171e193.
[26] J. Song, Z.-H. Wang, Impacts of mesic and xeric urban vegetation on outdoor
thermal comfort and microclimate in Phoenix, AZ, Build. Environ. 94 (2015)
558e568.
[27] M.J. Best, Representing urban areas within operational numerical weather
prediction models, Bound. Layer Meteorol. 114 (1) (2005) 91e109.
[28] F. Chen, H. Kusaka, R. Bornstein, J. Ching, C.S.B. Grimmond, S. GrossmanClarke, et al., The integrated WRF/urban modelling system: development,
evaluation, and applications to urban environmental problems, Int. J. Climatol.
31 (2) (2011) 273e288.
[29] A. Martilli, A. Clappier, M. Rotach, An urban surface exchange parameterisation for mesoscale models, Bound. Layer Meteorol. 104 (2) (2002) 261e304.
[30] F. Salamanca, A. Martilli, M. Tewari, F. Chen, A study of the urban boundary
layer using different urban parameterizations and high-resolution urban
canopy parameters with WRF, J. Appl. Meteorol. Climatol. 50 (5) (2011)
1107e1128.
[31] S. Miao, F. Chen, Formation of horizontal convective rolls in urban areas,
Atmos. Res. 89 (3) (2008) 298e304.
[32] H. Kusaka, F. Chen, M. Tewari, J. Dudhia, D.O. Gill, M.G. Duda, et al., Numerical
simulation of Urban Heat Island effect by the WRF Model with 4-km grid

J. Yang et al. / Building and Environment 108 (2016) 110e121

[33]

[34]

[35]

[36]
[37]

[38]

increment: an inter-comparison study between the urban canopy model and
slab model, J. Meteorol. Soc. Jpn. B 90 (2012) 33e45.
J.A. Fry, G. Xian, S. Jin, J.A. Dewitz, C.G. Homer, Y. LIMIN, et al., Completion of
the 2006 national land cover database for the conterminous United States,
Photogramm. Eng. Remote Sens. 77 (9) (2011) 858e864.
F. Ali-Toudert, H. Mayer, Numerical study on the effects of aspect ratio and
orientation of an urban street canyon on outdoor thermal comfort in hot and
dry climate, Build. Environ. 41 (2) (2006) 94e108.
J. Song, Z.-H. Wang, Interfacing the urban landeatmosphere system through
coupled urban canopy and atmospheric models, Bound. Layer Meteorol.
(2014) 1e22.
J. Yang, Z.-H. Wang, Optimizing urban irrigation schemes for the trade-off
between energy and water consumption, Energy Build. 107 (2015) 335e344.
J. Yang, Z.-H. Wang, F. Chen, S. Miao, M. Tewari, J. Voogt, et al., Enhancing
hydrologic modelling in the coupled weather research and forecastingeurban
modelling system, Bound. Layer Meteorol. 155 (1) (2015) 87e109.
U.S. Department of Energy. Energy Saver: Thermostats. http://energy.gov/

121

energysaver/thermostats. Retrieved Aug 8, 2016.
[39] J. Yang, Z.-H. Wang, K.E. Kaloush, Environmental impacts of reﬂective materials: is high albedo a ‘silver bullet’for mitigating urban heat island? Renew.
Sustain. Energy Rev. 47 (2015) 830e843.
[40] A. Brazel, N. Selover, R. Vose, G. Heisler, The tale of two climates-Baltimore
and Phoenix urban LTER sites, Clim. Res. 15 (2) (2000) 123e135.
[41] J.A. Patz, D. Campbell-Lendrum, T. Holloway, J.A. Foley, Impact of regional
climate change on human health, Nature 438 (7066) (2005) 310e317.
[42] A.K. Mishra, M. Ramgopal, Field studies on human thermal comfortdan
overview, Build. Environ. 64 (2013) 94e106.
[43] Givoni B. Estimation of the Effect of Climate on Man: Development of a New
Thermal Index: Technion-IIT, Building Research Station, 1963.
[44] D. Pearlmutter, P. Berliner, E. Shaviv, Physical modeling of pedestrian energy
exchange within the urban canopy, Build. Environ. 41 (6) (2006) 783e795.
[45] H. Akbari, S. Menon, A. Rosenfeld, Global cooling: increasing world-wide urban albedos to offset CO2, Clim. Change 94 (3e4) (2009) 275e286.