Solar Energy Vol. 70, No. 3, pp. 295–310, 2001
Published by Elsevier Science Ltd
Pergamon PII: S0038–092X(00)00089–X Printed in Great Britain
0038-092X/01 /$ - see front matter
COOL SURFACES AND SHADE TREES TO REDUCE ENERGY USE AND
IMPROVE AIR QUALITY IN URBAN AREAS
H. AKBARI , M. POMERANTZ and H. TAHA
Lawrence Berkeley National Laboratory, Heat Island Group, Berkeley, CA, USA
Abstract—Elevated summertime temperatures in urban ‘heat islands’ increase cooling-energy use and
accelerate the formation of urban smog. Except in the city’s core areas, summer heat islands are created mainly
by the lack of vegetation and by the high solar radiation absorptance by urban surfaces. Analysis of
temperature trends for the last 100 years in several large U.S. cities indicate that, since |1940, temperatures in
urban areas have increased by about 0.5–3.08C. Typically, electricity demand in cities increases by 2–4% for
each 18C increase in temperature. Hence, we estimate that 5–10% of the current urban electricity demand is
spent to cool buildings just to compensate for the increased 0.5–3.08C in urban temperatures. Downtown Los
Angeles (L.A.), for example, is now 2.58C warmer than in 1920, leading to an increase in electricity demand of
1500 MW. In L.A., smoggy episodes are absent below about 218C, but smog becomes unacceptable by 328C.
Because of the heat-island effects, a rise in temperature can have signiﬁcant impacts. Urban trees and
high-albedo surfaces can offset or reverse the heat-island effect. Mitigation of urban heat islands can
potentially reduce national energy use in air conditioning by 20% and save over $10B per year in energy use
and improvement in urban air quality. The albedo of a city may be increased at minimal cost if high-albedo
surfaces are chosen to replace darker materials during routine maintenance of roofs and roads. Incentive
programs, product labeling, and standards could promote the use of high-albedo materials for buildings and
roads. Similar incentive-based programs need to be developed for urban trees. Published by Elsevier Science
1. INTRODUCTION cities were cooler than their surroundings. After
1940, when built-up areas began to replace vege-
Modern urban areas have typically darker surfaces tation, the urban centers became as warm or
and less vegetation than their surroundings. These warmer than the suburbs, and the warming trend
differences affect climate, energy use, and became quite obvious, so that, from 1965 to 1989,
habitability of cities. At the building scale, dark urban temperatures have increased by about 18C.
roofs heat up more and, thus, raise the summer- Regardless of whether or not there is a tem-
time cooling demands of buildings. Collectively, perature difference from rural conditions, data
dark surfaces and reduced vegetation warm the air suggest that temperatures in cities are increasing.
over urban areas, leading to the creation of urban Fig. 1 depicts the summertime monthly maximum
‘heat islands’. On a clear summer afternoon, the and minimum temperatures between 1877 and
air temperature in a typical city is as much as 1997 in downtown Los Angeles. It clearly indi-
2.58C higher than in the surrounding rural areas. cates that the maximum temperatures at dow-
We have found that peak urban electric demand ntown Los Angeles are now about 2.58C higher
rises by 2–4% for each 18C rise in daily maxi- than they were in 1920. The minimum tempera-
mum temperature above a threshold of 15 to tures are about 48C higher than they were in 1880.
208C. Thus, the additional air-conditioning use In Washington, DC, temperatures increased by
caused by this urban air temperature increase is about 28C between 1871 and 1987. The data
responsible for 5–10% of urban peak electric indicate that this recent warming trend is typical
demand, at a direct cost of several billion dollars of most U.S. metropolitan areas, and exacerbates
annually. demand for energy.
In California, Goodridge (1987, 1989) showed Akbari et al. (1992) have found that peak urban
that, before 1940, the average urban–rural tem- electric demand in six American cities (Los
perature differences for 31 urban and 31 rural Angeles, CA; Washington, DC; Phoenix, AZ;
stations in California were always negative, i.e., Tucson, AZ; and Colorado Springs, CO) rises by
2–4% for each 18C rise in daily maximum
temperature above a threshold of 15 to 208C (the
Author to whom correspondence should be addressed. Tel.: case of Los Angeles is shown in Fig. 2). For the
11-510-486-4287; fax: 11-510-486-4673; e-mail:
h firstname.lastname@example.org Los Angeles Basin, it is estimated that the heat
296 H. Akbari et al.
Fig. 2 also shows the probability of smoggy days
in Los Angeles, as measured by ozone concen-
tration vs. temperature. At maximum daily tem-
peratures below 228C, the maximum concentra-
tion of ozone is below the California standard of
90 parts per billion (ppb); at temperatures above
358C, practically all days are smoggy.
2. HEAT ISLAND MITIGATION
Use of high-albedo urban surfaces and the
planting of urban trees are inexpensive measures
that can reduce summertime temperatures. The
effects of modifying the urban environment by
planting trees and increasing albedo are best
quantiﬁed in terms of ‘direct’ and ‘indirect’
contributions. The direct effect of planting trees
Fig. 1. Ten-year running average summertime monthly maxi- around a building or using reﬂective materials on
mum and minimum temperatures in Los Angeles, California roofs or walls is to alter the energy balance and
(1877–1997). The ten-year running average is calculated as cooling requirements of that particular building.
the average temperature of the previous four years, the current However, when trees are planted and albedo is
year, and the next ﬁve years. Note that the maximum modiﬁed throughout an entire city, the energy
temperatures have increased by about 2.58C since 1920. balance of the whole city is modiﬁed, producing
city-wide changes in climate. Phenomena associ-
island increases power consumption by about 1– ated with city-wide changes in climate are re-
1.5GW, costing the rate-payers over $100 million ferred to as indirect effects, because they indirect-
per year. Nationwide, the additional air-condition- ly affect the energy use in an individual building.
ing use caused by urban air temperature increase Direct effects give immediate beneﬁts to the
is responsible for 5–10% of urban peak electric building that applies them. Indirect effects achieve
demand, at a direct cost of several billion dollars beneﬁts only with widespread deployment.
annually. There is an important distinction between direct
Not only do summer heat islands increase and indirect effects: while direct effects are
system-wide cooling loads, but they also increase recognized and accounted for in present models of
smog production because of higher urban air building-energy use, indirect effects are ap-
temperatures (Taha et al
, 1994). For example, preciated far less. Accounting for indirect effects
is more difﬁcult and the results are comparatively
less certain. Understanding these effects and
incorporating them into accounts of energy use
and air quality is the focus of our current research.
It is worth noting that the phenomenon of summer
urban heat islands is itself an indirect effect of
The issue of direct and indirect effects also
enters into our discussion of atmospheric pollu-
tants. Planting trees has the direct effect of
reducing atmospheric CO because each individ-
ual tree directly sequesters carbon from the at-
mosphere through photosynthesis. However,
planting trees in cities also has an indirect effect
When sunlight hits an opaque surface, some of the energy is
reﬂected (this fraction is called the albedo 5a), and the
Fig. 2. Ozone levels and peak power for Southern California rest is absorbed (the absorbed fraction is 1 2a). Low-a
Edison versus 4 p.m. temperature in Los Angeles, California. surfaces of course become much hotter than high-a sur-
(Source: Akbari et al
, 1990). faces.
Cool surfaces and shade trees to reduce energy use and improve air quality in urban areas 297
on CO . By reducing the demand for cooling meteorology and its sensitivity to changes in
energy, urban trees indirectly reduce emission of surface properties. The Urban Airshed Model
CO from power plants. Akbari et al
(1990) (UAM) was used to simulate the impacts of the
showed that the amount of CO avoided via the changes in meteorology and emissions on ozone
indirect effect is considerably greater than the air quality. The CSUMM and the UAM essential-
amount sequestered directly. Similarly, trees di- ly solve a set of coupled governing equations
rectly trap ozone precursors (by dry-deposition), a representing the conservation of mass (continui-
direct effect, and indirectly reduce the emission of ty), potential temperature (heat), momentum,
these precursors from power plants (Taha, 1996). water vapor, and chemical species continuity to
obtain for prognostic meteorological ﬁelds and
pollutant species concentrations.
3. TOOLS FOR ANALYSIS The CSUMM is a hydrostatic, primitive-equa-
Fig. 3 depicts the overall methodology used in tion, three-dimensional Eulerian model that was
analyzing the impact of heat-island mitigation originally developed by Pielke (1974). The model
measures on energy use and urban air pollution. is incompressible (uses incompressibility assump-
The DOE-2 building-energy simulation program tion to simplify the equation for conservation of
is used to calculate the energy use and energy mass), and employs a terrain-following coordinate
savings in buildings. To calculate the direct system. It uses a ﬁrst order closure scheme in
effects, prototypical buildings are simulated with treating sub-grid scale terms of the governing
dark- and light-colored roofs, and with and with- differential equations. The model’s domain is
out shade trees. Typical weather data for each about 10 km high with an underlying soil layer
climate region of interest are used in these that is about 50 cm deep. The CSUMM generates
calculations. To calculate the indirect effects, the three-dimensional ﬁelds of prognostic variables as
typical weather data input to DOE-2 are ﬁrst well as a boundary layer height proﬁle that can be
modiﬁed to account for changes in the urban input to the UAM.
climate. The prototypical buildings are then simu- The UAM is a three-dimensional, Eulerian,
lated with the modiﬁed weather data to estimate photochemical model that is capable of simulating
savings in heating and cooling energy consump- inert and chemically reactive atmospheric pollu-
tion. tants. It has been recommended by the U. S.
To understand the impacts of large-scale in- Environmental Protection Agency (EPA) for
creases in albedo and vegetation on urban climate ozone air quality modeling studies of urban areas
and ozone air quality, mesoscale meteorological (EPA, 1986). The UAM simulates the advection,
and photochemical models are used. For example, diffusion, transformation, emission, and deposi-
Taha et al
(1995) and Taha (1996, 1997) used tion of pollutants. It treats about 30 chemical
the Colorado State University Mesoscale Model species and uses the carbon bond CB-IV mecha-
(CSUMM) to simulate the Los Angels Basin’s nism (Gery et al
, 1988). The UAM accounts for
Fig. 3. Methodology to analyze the impact of shade trees, cool roofs, and cool pavements on energy use and air quality (smog).
298 H. Akbari et al.
emissions from area and point sources, elevated 4. COOL ROOFS
stacks, mobile and stationary sources, and vegeta- At the building scale, a dark roof is heated by
tion (biogenic emissions). For a detailed discus- the sun and, thus, directly raises the summertime
sion of the use and adaptation of these models and cooling demand of the building beneath it. For
the study of the impact of the heat island mitiga- highly absorptive (low-albedo) roofs, the differ-
tion strategies in L.A. Basin, see Taha (1996, ence between the surface and ambient air tem-
1997). peratures may be as high as 508C, while for less
Examples of outputs from these simulations are absorptive (high-albedo) surfaces with similar
shown in Figs. 4 and 5. Fig. 4 shows the predicted insulative properties, such as roofs covered with a
reduction in air temperature in Los Angeles at 2 white coating, the difference is only about 108C
p.m. on August 27 as a result of increasing the (Berdahl and Bretz, 1997). For this reason, ‘cool’
urban albedo and vegetation cover by moderate surfaces (which absorb little ‘insolation’) can be
amounts (average increases of 7%). Fig. 5 shows effective in reducing cooling-energy use. Highly
corresponding changes in ozone concentrations. absorptive surfaces contribute to the heating of
Because of the combined effects of local emis- the air, and thus indirectly increase the cooling
sions, meteorology, surface properties, and topog- demand of (in principle) all buildings. Cool
raphy, ozone concentrations increase in some surfaces incur no additional cost if color changes
areas and decrease in others. The net effect, are incorporated into routine re-rooﬁng and re-
however, is a decrease in ozone concentrations. surfacing schedules (Bretz et al
, 1997 and Rosen-
The simulations also predict a reduction in feld et al
population-weighted exceedance exposure to Most high-albedo surfaces are light colored,
ozone (above the California and National Ambient although selective surfaces that reﬂect a large
Air Quality Standards) of 10–20% (Taha, 1996). portion of the infrared solar radiation but absorb
This reduction, for some smog scenarios, is some visible light may be dark colored and yet
comparable to ozone reductions obtained by have relatively high albedos (Berdahl and Bretz,
replacing all gasoline on-road motor vehicles with 1997).
Fig. 4. Temperature difference (from the base case) for a case with increased surface albedo and urban forest. The temperature
difference is at 2 p.m. on a late-August day in Los Angeles.
Cool surfaces and shade trees to reduce energy use and improve air quality in urban areas 299
Fig. 5. Ozone concentrations difference (from the base case) for a case with increased surface albedo and urban forest. The
difference is shown for 2 p.m. on a late-August day in Los Angeles.
Energy and smog beneﬁts of cool roofs modeled with the DOE-2.1E simulation program.
Akbari et al
(1993) and Gartland et al
Direct energy savings. There is a sizable report that the simulations underestimated the
body of measured data documenting the direct cooling-energy savings and peak-power reduc-
energy-saving effects of light-colored roofs. In the tions by as much as twofold.
summers of 1991 and 1992, Akbari et al
(1993, Parker et al
(1995) monitored nine homes in
1997) monitored peak power and cooling-energy Florida before and after applying high-albedo
savings from high-albedo coatings on one house coatings to their roofs. Air-conditioning energy
and two school bungalows in Sacramento, Cali- use was reduced by 10–43%, with average sav-
fornia. They collected data on air-conditioning ings of 7.4 kWh/day (savings of 19%). Peak
electricity use, indoor and outdoor temperatures demand between 5 and 6 p.m. was reduced by
and humidities, roof and ceiling surface tempera- 0.2–1.0 kW, with an average reduction of 0.4 kW
tures, inside and outside wall temperatures, inso- (savings of 22%). The amount of energy savings
lation, and wind speed and direction. roughly inversely correlated with the amount of
Applying a high-albedo coating to one house ceiling insulation and the location of the duct
resulted in seasonal savings of 2.2 kWh/day (80% system: large savings in poorly insulated homes
of base-case use), and peak demand reductions of and those with the duct systems in the attic space,
0.6 kW (about 25% of base-case demand). In the and smaller savings in well insulated homes.
school bungalows, cooling-energy was reduced by Akbari et al
(1998) and Konopacki et al
3.1 kWh/day (35% of base-case use), and peak (1998) monitored the impacts of light-colored
demand by 0.6 kW (about 20% of base-case roofs on cooling-energy use of three commercial
demand). (It is important to note that altering the buildings in northern California. Increasing the
albedo starts to pay for itself immediately through reﬂectance of the roofs from an initial albedo of
the direct effect.) The buildings were also about 0.20 to 0.60 dropped the roof temperature
300 H. Akbari et al.
on hot summer afternoons by about 258C. Sum- Angeles, Dallas/Fort Worth, Houston, Miami/
mertime, standard-weekday, average daily air- Fort Lauderdale, New Orleans, New York City,
conditioning savings were 18% in a medical Philadelphia, Phoenix, and Washington, DC/
ofﬁce building, 13% in a second medical ofﬁce Baltimore. Cooling-energy savings and heating-
building, and 2% in a drug store. In another energy penalties were then obtained from the
demonstration project in Florida, Parker et al
difference in the simulated energy use of the
(1998) measured cooling electricity savings re- prototype buildings with light- and dark-colored
sulting from the application of light-colored rooﬁ- roofs.
ng in a small strip mall; they reported savings of The study also estimated how much energy and
about 20 to 40%. The Sacramento Municipal money could be saved if all the roofs of existing
Utility District (SMUD) reports similar savings, building stocks in large metropolitan areas were
measured in about ten commercial buildings in changed from dark to light. This was done by
Sacramento (Hildebrandt et al
, 1998). scaling the simulated energy savings of the in-
Computer simulations are used to obtain esti- dividual prototype buildings by the amount of
mates of year-round effects for a variety of air-conditioned space immediately beneath all
building types and climates. A recent study made roofs in an entire MSA. For this purpose, we used
quantitative estimates of peak demand and annual data on the stock of commercial and residential
cooling-electricity use and savings that would buildings in each MSA, the saturation of heating
result from increasing the reﬂectivity of the roofs and cooling systems, the current roof reﬂec-
(Konopacki et al
, 1997). The estimates of annual tivities, and the local costs of electricity and gas.
net savings in cooling electricity are adjusted for Results for the 11 metropolitan areas are sum-
the penalty of increased wintertime heating- marized in Tables 1 and 2. Sum totals for all 11
energy use. The analysis is based on simulation of MSAs were: electricity savings, 2.6 tera-watt
building-energy use, using the DOE-2 building- hours (TWh) (200 kilowatt hours per 100 m roof
energy simulation program. The study speciﬁed area of air-conditioned buildings); heating energy
11 prototypical buildings: single-family residen- penalty, 6.9 TBtu (5 therms per 100 m ); net
tial (old and new), ofﬁce (old and new), retail savings in energy bills, $194 M ($15 per 100 m );
store (old and new), school (primary and sec- and savings in peak demand 1.7 gigawatt (GW)
ondary), health care (hospital and nursing home), (135 W per 100 m ). Six building types account
and grocery store. Most prototypes were simu- for over 90% of the annual electricity and net
lated with two heating systems: gas furnace and energy savings: old residences accounted for more
heat pumps. DOE-2 simulations were performed than 55%, new residences for about 15%, and
for the prototypical buildings, with light and dark four other building types (old/new ofﬁces and
roofs, in a variety of climates, to obtain estimates
of the energy use for air conditioning and heating.
Weather data for 11 U.S. Metropolitan Statistical
Areas (MSAs) were used: Atlanta, Chicago, Los One therm is 100,000 Btu.
Table 1. Estimates of metropolitan-scale annual cooling electricity savings (GWh), net energy savings ($M), peak demand
electricity savings (MW), and annual natural gas penalty (GBtu) resulting from application of light-colored rooﬁng on residential
and commercial buildings in 11 Metropolitan Statistical Areas
Metropolitan area Residential Commercial Commercial and residential
Elec Gas Net Peak Elec Gas Net Peak Elec Gas Net Peak
(GWh) (GBtu) (M$) (MW) (GWh) (GBtu) (M$) (MW) (GWh) (GBtu) (M$) (MW)
Atlanta 125 349 8 83 22 55 1 14 147 404 9 97
Chicago 100 988 6 89 84 535 4 56 183 1523 10 145
Los Angeles 210 471 18 218 209 154 18 102 419 625 35 320
Dallas/Ft Worth 241 479 16 175 71 113 4 36 312 592 20 211
Houston 243 284 21 127 79 62 6 30 322 347 27 156
Miami/Ft Lauderdale 221 4 18 115 35 3 2 11 256 7 20 125
New Orleans 84 107 6 27 33 28 3 16 117 135 9 42
New York 35 331 3 56 131 540 13 95 166 871 16 151
Philadelphia 44 954 21 108 47 292 4 49 91 1246 3 157
Phoenix 299 74 32 106 58 31 5 18 357 105 37 123
DC/Baltimore 182 845 6 183 45 184 2 31 227 1029 8 214
Total 1784 4886 133 1287 814 1997 62 458 2597 6884 194 1741
Cool surfaces and shade trees to reduce energy use and improve air quality in urban areas 301
Table 2. Estimates of savings or penalties per 100 m of roof area of air-conditioned buildings resulting from application of
light-colored rooﬁng on residential and commercial buildings in 11 Metropolitan Statistical Areas: annual cooling electricity
savings (kWh), net energy savings ($), peak demand electricity savings (W), and annual natural gas penalty (therms)
Metropolitan area Residential Commercial and residential Commercial
Elec Gas Net Peak Elec Gas Net Peak Elec Gas Net Peak
(kWh) (therms) ($) (W) (kWh) (therms) ($) (W) (kWh) (therms) ($) (W)
Atlanta 153 4 10 102 239 6 11 152 162 4 10 107
Chicago 131 13 8 116 228 15 11 152 162 13 9 128
Los Angeles 182 4 16 189 350 3 30 171 239 4 20 183
Dallas/Ft Worth 166 3 11 121 224 4 13 114 176 3 11 119
Houston 198 2 17 103 261 2 20 99 211 2 18 102
Miami/Ft Lauderdale 259 0 21 135 340 0 19 107 267 0 21 131
New Orleans 199 3 14 64 287 2 26 139 218 3 17 78
New York 104 10 9 166 211 9 21 153 173 9 17 158
Philadelphia 81 18 22 199 232 14 20 241 122 17 4 211
Phoenix 314 1 34 111 409 2 35 127 327 1 34 113
DC/Baltimore 137 6 5 138 221 9 10 152 148 7 5 140
old/new retail stores) together accounted for tential savings were valued at $104M /year
(Rosenfeld et al
The results for the 11 MSAs were extrapolated
Other beneﬁts of cool roofs
to estimate the savings in the entire United States. Another beneﬁt of a light-colored roof is a
The study estimates that, nationally, light-colored potential increase in its useful life. The diurnal
rooﬁng could produce savings of about 10 TWh/ temperature ﬂuctuation and concomitant expan-
year (about 3.0% of the national cooling-electrici- sion and contraction of a light-colored roof is
ty use in residential and commercial buildings), an smaller than that of a dark one. Also, the degra-
increase in natural gas use by 26 GBtu/year dation of materials due to absorption of ultra-
(1.6%), a decrease in peak electrical demand of violet light is a temperature-dependent process.
7 GW (2.5%) (equivalent to 14 power plants each For these reasons, cooler roofs may last longer
with a capacity of 0.5 GW), and a decrease in net than hot roofs of the same material.
annual energy bills for the rate-payers of $750M.
Potential problems with cool roofs
Indirect energy and smog beneﬁts. Using
the Los Angeles Basin as a case study, Taha Several possible problems may arise from the
(1996, 1997) examined the impacts of using cool use of reﬂective rooﬁng materials (Bretz and
surfaces (cool roofs and pavements) on urban air Akbari, 1994, 1997). A drastic increase in the
temperature and, thus, on cooling-energy use and overall albedo of the many roofs in a city has the
smog. If higher albedo surfaces are thoroughly potential to create glare and visual discomfort if
applied, an urban heat island can be limited or not kept to a reasonable level. Besides being
reversed at negligible expense. In these simula- unpleasant, extreme glare could possibly increase
tions, Taha estimates that about 50% of the the incidence of trafﬁc accidents. Fortunately, the
urbanized area in the L.A. Basin is covered by glare for ﬂat roofs is not a major problem for
roofs and roads, the albedos of which can realisti- those who are at street level. For sloped roofs, the
cally be raised by 0.30 when they undergo normal problem of glare should be studied in detail
repairs. This results in a 28C cooling at 3 p.m. before proceeding with a full-scale implementa-
during an August episode. This summertime tion of this measure.
temperature reduction has a signiﬁcant effect on In addition, many types of building materials,
further reducing building cooling-energy use. The such as tar rooﬁng, are not well adapted to
annual savings in L.A. are estimated at $21M painting. Although such materials could be spe-
(Rosenfeld et al
, 1998). cially designed to have a higher albedo, this
Taha has also simulated the impact of urban- would be at a greater expense than painting.
wide cooling in Los Angeles on smog; the results Additionally, to maintain a high albedo, roofs
show a signiﬁcant reduction in ozone concen- may need to be recoated or rewashed on a regular
tration. The simulations predict a reduction of basis. The cost of a regular maintenance program
10–20% in population-weighted smog (ozone). In could be signiﬁcant.
L.A., where smog is especially serious, the po- A possible conﬂict of great concern is the fact
302 H. Akbari et al.
that building owners and architects like to have shade air-conditioning equipment would also like-
the choice as to what color to select for their ly be beneﬁcial.
rooftops. This is particularly a concern for sloped
Energy and smog beneﬁts of shade trees
Direct energy savings. Data on mea-
Cost of cool roofs sured energy savings from urban trees are scarce.
Increasing the overall albedo of roofs is an In one experiment, Parker (1981) measured the
attractive way of reducing the net radiative heat cooling-energy consumption of a temporary build-
gains through the roof, and, hence, reducing ing in Florida before and after adding trees and
building cooling loads. To change the albedo, the shrubs and found cooling-electricity savings of up
rooftops of buildings may be painted or covered to 50%. In the summer of 1992, Akbari et al
with a new material. Since most roofs have (1997) monitored peak-power and cooling-energy
regular maintenance schedules or need to be re- savings from shade trees in two houses in Sac-
roofed or recoated periodically, the change in ramento, California. The collected data included
albedo should be done then to minimize the costs. air-conditioning electricity use, indoor and out-
High-albedo alternatives to conventional rooﬁ- door dry-bulb temperatures and humidities, roof
ng materials are usually available, often at little or and ceiling surface temperatures, inside and out-
no additional cost. For example, a built-up roof side wall temperatures, insolation, and wind speed
typically has a coating or a protective layer of and direction. The shading and microclimate
mineral granules or gravel. Under such condi- effects of the trees at the two monitored houses
tions, it is expected that choosing a reﬂective yielded seasonal cooling-energy savings of 30%,
material at the time of installation should not add corresponding to average savings of 3.6 and 4.8
to the cost of the roof. Also, rooﬁng shingles are kWh/day. Peak-demand savings for the same
available in a variety of colors, including white, at houses were 0.6 and 0.8 kW (about 27% savings
the same price. The incremental price premium in one house and 42% in the other).
for choosing a white rather than a black single-ply A few other studies have focused on the wind-
membrane rooﬁng material is less than 10%. Cool shielding effect of trees. DeWalle et al
rooﬁng materials that require an initial investment used mobile homes to measure the windbreaking
may turn out to be more attractive in terms of effects of trees on energy use. In a follow-up
life-cycle cost than conventional dark alternatives. experiment, Heisler (1989) measured the effect of
Usually, the lower life-cycle cost results from trees on wind and solar radiation in a residential
longer roof life and/or energy savings. neighborhood. Huang et al
(1990) used the data
provided by Heisler (1989) and simulated the
impact of shading and wind reduction on residen-
5. URBAN TREES tial buildings’ heating- and cooling-energy use.
The beneﬁts of trees can also be divided into Their simulations indicated that a reduction in
direct and indirect effects: shading of buildings inﬁltration because of trees would save heating-
and ambient cooling (urban forest). Shade trees energy use. However, in climates with cooling-
intercept sunlight before it warms a building. The energy demand, the impact of windbreak on
urban forest cools the air by evapotranspiration. cooling is fairly small compared to the shading
Trees also decrease the wind speed under their effects of trees and, depending on climate, it
canopy and shield buildings from cold winter could decrease or increase cooling-energy use. In
breezes. Urban shade trees offer signiﬁcant bene- cold climates, the wind-shielding effect of trees
ﬁts by both reducing building air-conditioning, can substantially reduce heat-energy use in build-
lowering air temperature, and thus improving ings. Akbari and Taha (1992) simulated the wind-
urban air quality by reducing smog. Over the life shielding impact of trees on heating-energy use in
of a tree, the savings associated with these four Canadian cities. For several prototypical
beneﬁts vary by climate region and can be up to residential buildings, they estimated heating-
$200 per tree. The cost of planting trees and energy savings in the range of 10 to 15%.
maintaining them can vary from $10 to $500 per In a recent study, Taha et al
tree. Tree-planting programs can be designed to the meteorological impact of large-scale tree-
be low cost, so they can offer savings to com- planting programs in ten U.S. metropolitan areas:
munities that plant trees. We are considering here Atlanta, GA; Chicago, IL; Dallas, TX; Houston,
trees that shade buildings. Placing trees in order to TX; Los Angeles, CA; Miami, FL; New York,
Cool surfaces and shade trees to reduce energy use and improve air quality in urban areas 303
Table 3. Number of additional trees planted in each metropolitan area and their simulated effects in reducing the ambient
temperature. (Source: Taha et al
Location Millions of Millions of Max air temperature
addional trees additional trees reduction in the
in the simulation in the metropolitan hottest simulation
domain area cell (8C)
Atlanta 3.0 1.5 1.7
Chicago 12 5.0 1.4
Los Angeles 11 5.0 3.0
Fort Worth 5.6 2.8 1.6
Houston 5.7 2.7 1.4
Miami 3.3 1.3 1.0
New York City 20 4.0 2.0
Philadelphia 18 3.8 1.8
Phoenix 2.8 1.4 1.4
Washington, DC 11 3.0 1.9
NY; Philadelphia, PA; Phoenix, AZ; and simulation cells, the temperature was decreased
Washington, DC). Table 3 shows the number of by up to 38C (see Table 3). The corresponding
added trees simulated in each city and impact on air-conditioning savings resulting from ambient
air temperature. The number of trees in each grid cooling by trees in hot climates ranges from $5 to
cell varied from the low hundreds to the high tens $10 per year per 100 m of roof area of residential
of thousands. The DOE-2 building simulation and commercial buildings. Indirect effects are
program was then used to estimate the direct and smaller than direct shading, and, moreover, re-
indirect impacts of trees on saving cooling-energy quire that the entire city be planted.
use for two building prototypes: a single-family Based on the results of Taha et al
residence and an ofﬁce. The calculations ac- Rosenfeld et al
(1998) estimated the potential
counted for a potential increase in winter heating- beneﬁts of trees, speciﬁcally in the Los Angeles
energy use. Table 4 shows that, in most hot cities, Basin. The study assumed planting 11M trees
shading a building can save annually $5 to $25 according to the following plan: three shade trees
per 100 m of roof area of residential and (each with a canopy cross section of 50 m ) per
commercial buildings. Savings in residential air-conditioned house, for a total of 5.4M trees;
building are higher than in commercial buildings. about one shade tree for each 250 m of non-
residential roof area for a total of 1M trees; 4.6M
Indirect energy and smog beneﬁts. Taha trees to shade non-air-conditioned homes or to be
(1996) estimated the impact on ambient planted along streets, in parks, and in other public
temperature resulting from a large-scale tree- spaces. The results of that analysis are shown in
planting program in the selected ten cities. They Table 5. Note that about two-thirds of the savings
used a three-dimensional meteorological model to in L.A. result from the reduction in smog con-
simulate the potential impact of trees on ambient centration resulting from meteorological changes
temperature for each region. The mesoscale simu- due to the evapotranspiration of trees. It has also
lations showed that, on average, trees can cool been suggested that trees improve air quality by
down cities by about 0.3 to 18C at 2 p.m.; in some dry-depositing NO , O , and PM10 particulates.
Table 4. DOE-2 simulated HVAC annual energy savings from trees. Three trees per house and per ofﬁce are assumed. All
savings are $/100 m . (Source: Taha et al
Location Old residence New residence Old ofﬁce New ofﬁce
Direct Indirect Indirect Direct Indirect Direct Indirect Direct
Atlanta 5 2 3 1 3 2 2 2
Chicago 3 2 1 0.5 1 1 2 1
Los Angeles 12 8 7 5 6 12 4 10
Fort Worth 6 6 5 4 4 5 2 4
Houston 10 6 6 4 3 5 3 3
Miami 9 3 6 3 3 2 2 2
New York City 3 2 2 1 3 3 2 2
Philadelphia 250 2702111
Phoenix 27 8 16 5 9 5 6 4
Washington, DC 3 2 1 1 3 1 2 1
304 H. Akbari et al.
Table 5. Energy savings, ozone reduction, and avoided peak power resulting from use of urban trees in the Los Angeles Basin
(Source: Rosenfeld et al
Beneﬁts Direct Indirect Smog Total
1 Cost savings from trees (M$/ year) 58 35 180 273
2DPeak power (GW) 0.6 0.3 0.9
3 Present value per tree ($) 68 24 123 211
Rosenfeld et al
(1998) estimate that 11M trees in other beneﬁts of trees are not considered in the
L.A. will reduce PM10 by less than 0.1%, worth cost beneﬁt analysis shown in this paper.
only $7M, which is disappointingly smaller than
Potential problems with shade trees
the beneﬁts of $180M from smog reduction. There are some potential problems associated
The present value (PV) of savings is calculated with trees. Some trees emit volatile organic
to ﬁnd out how much a homeowner can afford to compounds (VOCs) that exacerbate the smog
pay for shade trees. Rosenfeld et al
(1998) problem. Obviously, selection of low-emitting
estimate that, on this basis, the direct savings to a trees should be considered in a large-scale tree-
home owner who plants three shade trees would planting program. Benjamin et al
have a present value of about $200 per home prepared a list of several hundred tree species
($68/tree). The present value of indirect savings with their average emission rates.
was smaller, about $72/home ($24/tree). The PV In dry climates and areas with a serious water
of smog savings was about $120/tree. Total PV of shortage, drought-resistant trees are recom-
all beneﬁts from trees was then $210/tree. mended. Some trees need signiﬁcant maintenance
Reducing smog by citywide cooling can be that may entail high cost over the life of the trees.
considered equivalent to reducing the formation Tree roots can damage underground pipes, pave-
of smog precursors at constant temperature. We ments and foundations. Proper design is needed to
estimate that shade trees will reduce the maxi- minimize these effects. Also, trees are a fuel
mum smog concentration by 5%. Using the ozone source for ﬁre; selection of appropriate tree
‘isopleths’ (such as Milford’s), a 5% reduction in species and planting them strategically to mini-
smog is equivalent to reducing precursors by mize the ﬁre hazard should be an integral com-
approximately 12%, i.e., reducing NO in L.A. by
ponent of a tree-planting program.
175 tons/day, a very signiﬁcant drop and 25 times
Cost of trees
more than the 4 tons/day through reduced power-
plant emissions. The cost of a citywide tree-planting program
depends on the type of program offered and the
Other beneﬁts of shade trees types of trees recommended. At the low end, a
There are other beneﬁts associated with urban promotional planting of trees 5–10 feet high costs
trees. Some of these include improvement in the about $10 per tree, whereas a professional tree-
quality of life, increased value of properties, planting program using fairly large trees could
decreased rain run-off water and, hence, a protec- amount to $150 to $470 a tree (McPherson et al
tion against ﬂoods (McPherson et al
, 1994). 1994). McPherson has collected data on the cost
Trees also directly sequester atmospheric carbon of tree planting and maintenance from several
dioxide, but Rosenfeld et al
(1998) estimate that cities. The cost elements include planting, prun-
the direct sequestration of carbon dioxide is less ing, removal of dead trees, stump removal, waste
than one-fourth of the emission reduction re- disposal, infrastructure repair, litigation and
sulting from savings in cooling-energy use. These liability, inspection, and program administration.
The data provide details of the cost for trees
located in parks, yards, along streets, highways,
and houses. The present value of all of these
Milford et al
(1989) have carried out detailed calculations
analyzing the changes in the maximum ozone concen- life-cycle costs (including planting) is $300 to
tration reached in Los Angeles vs. initial concentration of $500 per tree. Over 90% of the cost is associated
NO and VOCs (volatile organic compounds). They pre-
with professional planting, pruning, and tree and
sented their calculations in the form of ‘isopleths’ of equal stump removal. On the other hand, a program
maximum smog concentration for various levels of NO
administered by the Sacramento Municipal Utility
and VOCs concentration (typically shown as a percent
reduction of emissions) for a typical summer episode. District (SMUD) and Sacramento Tree Founda-
Cool surfaces and shade trees to reduce energy use and improve air quality in urban areas 305
tion in 1992–1996 planted 20-foot tall trees at an 6. COOL PAVEMENTS
average cost of $45 per tree. This only includes The practice of widespread paving of city
the cost of a tree and its planting; it does not streets with asphalt began only within the past
include pruning, removal of dead trees, and hundred years. The advantages of this smooth and
stump. With this wide range of costs associated all-weather surface for the movement of bicycles
with trees, in our opinion, tree costs should be and automobiles is obvious, but some of the
justiﬁed by other amenities they provide beyond associated problems are perhaps not so well
air-conditioning and smog beneﬁts. The best appreciated. One consequence of covering streets
programs are then probably the information pro- with dark asphalt surfaces is the increased heating
grams that provide data on energy and smog of the city by sunlight. A dark surface absorbs
savings of trees to the communities and home light, and, therefore, it gets warmer. The pave-
owners that have decided to plant trees for other ments in turn heat the air and help create the
reasons. ‘urban heat island’. If urban surfaces were lighter
Even trees planted along streets and in parks in color, more of the incoming light would be
where they do not offer direct shade to air-con- reﬂected back into space and the surfaces and the
ditioned buildings exert an ambient cooling effect air would be cooler. This tends to reduce the need
sufﬁcient to have a substantial impact on smog for air conditioning.
reduction. Simulations for Los Angeles indicate Urban pavements are made predominantly of
that trees account for net savings (energy and asphalt concrete. In this discussion, we will not
smog savings) of about $270M annual beneﬁt, of deal with the common alternative, cement con-
which, $58M comes from their contribution to crete, and the ongoing debate as to whether it is
shading (Table 5). preferable because of its longer life-time. The
At another level, our calculations suggest that questions we address are whether there are ways
urban trees play a major role in sequestering CO
to reduce the heating of cities caused by asphalt
and thereby delaying global warming. Rosenfeld
(1998) showed that a tree planted in Los concrete and whether this can be economical and
Angeles avoids the combustion of 18 kg of carbon practical.
annually, even though it sequesters only 4.5 kg (as In Fig. 6, we show some measurements of the
it would if growing in a forest). In that sense, one effect of albedo on pavement temperature. The
shade tree in Los Angeles is equivalent to four data clearly indicate that signiﬁcant modiﬁcation
forest trees. of the pavement temperature can be achieved: a
Fig. 6. Dependence of pavement surface temperature on albedo. Data were taken at about 3 p.m. in Berkeley, California, on new,
old, and light-color coated asphalt pavements. The data from San Ramon, California, were taken at about 3 p.m. on four asphalt
concrete and one cement concrete (albedo50.35) pavements.
306 H. Akbari et al.
108C decrease in temperature for a 0.25 increase ozone exceedance of the California air-quality
in albedo. standard (Taha, 1997). It has been estimated (Hall
, 1992) that L.A. people would be willing to
pay about $10 billion per year to avoid the
Energy and smog beneﬁts of cool medical costs and lost work time due to air
pavements pollution. The greater part of pollution is par-
An estimate of the beneﬁts can be deduced by ticulates, but the ozone contribution averages
ﬁrst ﬁnding the temperature decrease that would about $3 billion/year. Assuming a proportional
result if a city were resurfaced with more reﬂec- relationship of the cost with the amount of smog
tive paving materials. Cool pavements provide exceedance, the cooler-surfaced city would save
only indirect effects through lowered ambient 12% of $3 billion/year, or $360M/year. As
temperatures. Lower temperature has two effects: above, we attribute about 21% of the saving to
(1) reduced demand for electricity for air con- pavements. Thus, smog improvement from alter-
ditioning and (2) decreased production of smog ing the albedo of all 1250 km of pavements by
(ozone). Rosenfeld et al
(1998) estimated the cost 0.25 saves about $76M/year. Per unit area, this is
savings of reduced demand for electricity and of worth about $0.06/m per year.
the externalities of lower ozone concentrations in
the Los Angeles Basin.
Other beneﬁts of cool pavements
Electric power savings in Los Angeles.
Increased life expectancy of pavements.
Simulations for Los Angeles indicate that a It has long been known that the temperature of a
reasonable change in the albedo of the city could pavement affects its performance (Yoder and
cause a noticeable decrease in temperature. Taha Witzak, 1975). This has been emphasized by the
(1997) predicted a 1.58C decrease in temperature
new system of binder speciﬁcation advocated by
of the downtown area. The lower temperatures in the Strategic Highway Research Program
the city are calculated for the condition that all (SHRP). Beginning in 1987, this program led
roads and roofs are improved. From the pavement experts to carry out the task of re-
meteorological simulations of 3 days in each searching and then recommending the best meth-
season, the temperature changes for every day in a ods of making asphalt concrete pavements. A
typical year were estimated for Burbank, typical result of this study was the issuance of spe-
of the hottest one-third of L.A. The energy ciﬁcations for the asphalt binder. The temperature
consumptions of typical buildings were then range that the pavement will endure is a primary
simulated for the original weather and also for the consideration (Cominsky et al
, 1994). The per-
modiﬁed weather. The differences are the annual formance grade (PG) is speciﬁed by two tempera-
energy changes due to the decrease in ambient tures: (1) the average 7-day maximum tempera-
temperature. The result is a city-wide annual ture that the pavement will likely encounter, and
saving of about $71M (million), due to combined (2) the minimum temperature the pavement will
albedo and vegetation changes. The kWh savings likely attain. Note, importantly, that it is the
attributable to the pavement are $15M/year, or
pavement temperature and not the air temperature
$0.012/m year. Analysis of the hourly demand that is considered. There is a rule of thumb in the
indicates that cooler pavements could save an industry, the ‘rule of 90’, that, when the sum of
estimated 100 MW of peak power in L.A. Fewer the absolute values of these temperatures is
power plants, required to handle the growth of greater than 908C, some kind of modiﬁcation of
peak load, need be built if cooler pavements are the asphalt will be needed; this adds to the cost.
installed, saving money, resources, and pollution. For example, if a binder is speciﬁed as PG 58-22,
Smog savings in Los Angeles. The simu- it is intended to function between 58 and 2228C.
lations of the effects of higher albedo on smog The sum of the absolute values, 581u222u580.
formation indicate that an albedo change of 0.3 An ordinary grade of asphalt will sufﬁce; its cost
throughout the developed 25% of the city would is about $125 per ton. If, however, the pavement
yield a 12% decrease in the population-weighted must function between 76 and 2168C, or PG
76-16, the sum 761u216u592, a modiﬁed
asphalt is recommended. This higher grade costs
The model assumes that all roofs (1250 km ) have albedo about $165 per ton (Bally, 1998), a 30% increase
increased by 0.35, from about 0.15 to 0.5, and all
pavements (1250 km ) have albedo increased from about
0.1 to 0.35. As an example of a temperature-dependent
Cool surfaces and shade trees to reduce energy use and improve air quality in urban areas 307
source of pavement failure, we consider ‘tertiary hard and loosen the aggregate. Aggregate is
creep’. Pavements gradually accumulate perma- sometimes thrown by tires, but, when installed
nent distortions as tires roll over them repeatedly. properly, this seems to be rare.
‘Tertiary creep’ refers to the phenomenon of the
accelerated rate of distortion after many such
Cost of cool pavements
repetitions. This signals gross failure of the It is clear that cooler pavements will have
pavement. Cominsky et al
(1994) gave evidence energy, environmental and engineering beneﬁts.
that cooler pavements can have a signiﬁcantly The issue is then whether there are ways to
higher resistance to tertiary creep and, thus, can construct pavements that are feasible, economical,
last much longer. Using Cominsky’s data, we and cooler. The economic question is whether or
estimate that a 108C decrease in pavement tem- not the savings generated by a cool pavement
perature can result in a 25-fold increase in its life over its lifetime are greater than its extra cost.
expectancy against this type of failure, a factor so Properly, one should distinguish between initial
large that we suspect it is an overestimate. cost and lifetime costs (including maintenance,
repair time, and length of service of the road).
Improved visibility. Reﬂectivity of pave- Often the initial cost is decisive, so we will
ments is also a safety factor in visibility at night consider only that cost here.
and in wet weather, affecting the demand for
streets’ electric lighting. Street lighting is more
Thick pavements. A typical asphalt con-
effective if pavements are more reﬂective, which crete contains about 7% of asphalt by weight, or
can lead to greater safety; or, alternatively, less about 17% by volume; the remainder is rock
lighting could be used to obtain the same visibili- aggregate, except for a few percent of voids. The
ty. These beneﬁts have not yet been monetized. cost of ordinary asphalt (1998 prices) is about
$125 per ton, and the price of aggregate is about
Potential problems with cool pavements $20 per ton, exclusive of transportation costs.
A practical drawback of high reﬂectivity is Thus, in one ton of mixed asphalt concrete, the
glare, but this does not appear to be a problem. cost of materials only is about $28/ton, of which
We suggest a change in resurfacing using not about $9 is in the binder and $19 is in the
black asphalt, with an albedo of about 0.05–0.12, aggregate. For a pavement about 10 cm thick (4
but the application of a product with an albedo of inches), with a density of 2.1 ton/m , the cost of
about 0.35, similar to that of cement concrete. The the binder alone is about $2 per m and aggregate
experiment to test whether this will be a problem costs about $4.2 per m .
has already been performed: every day, millions Experimentally, the albedo of a fresh asphalt
of people drive on cement concrete roads, and we concrete pavement is about 0.05 (Pomerantz et
rarely hear of accidents caused by glare, or of al
, 1997) because the relatively small amount of
people even complaining about the glare on such black asphalt coats the lighter colored aggregate.
roads. Thus, every reader of this paper likely As an asphalt concrete pavement is worn down
knows the answer from experience. and the aggregate is revealed, we observed an
There is also a concern that, after some time, albedo increase to about 0.15 for ordinary aggre-
light-colored pavement will darken because of gate. If it were made with a reﬂective aggregate,
dirt. This tends to be true, but again, experience we could expect the long-term albedo to approach
with cement concrete roads suggests that the light that of the aggregate.
color of the pavement persists after long usage. How much money might such a reﬂective
Most drivers can see the difference in reﬂection pavement save? Using the assumptions for Los
between an asphalt and a cement concrete road Angeles, a cooler pavement would generate a
when they drive over them, even when the roads stream of savings of $0.07/m per year for the
are old. More studies are needed to quantify the lifetime of the road, about 20 years. At a real
effect of aging. interest rate of 3% per year, this has a present
The use of chip seals is a promising method of value about 15 times the current saving (Rosen-
resurfacing that achieves lighter color. (Chip feld et al
, 1998). Thus, the potential savings are
sealing is the pressing on of aggregate into soft worth $1.1/m at present. This saving would
binder, as a resurfacing technique (Asphalt Insti- allow for purchase of a binder, instead of $2/m ,
tute, 1989).) Although this is popular in many costing $3/m , or 50% more expensive. Or, one
districts, some have reservations about the use of could buy aggregate; instead of spending $4.2/m ,
chip seals in cul de sacs, where tires are turned one can now afford $5.2/m (a 20% more expen-
308 H. Akbari et al.
sive, whiter, aggregate). It is doubtful that such surface, achieved by using a chip seal. The costs
of both of these are about the same, $0.60/m
modest increases in costs can buy much whiter (Means, 1996). For a chip seal, about half the
pavements. materials cost is aggregate and half is the binder.
In the special case of a climate in which the If special light-colored aggregate is used in the
pavement is subjected to such wide temperature chip seal, there will be an extra cost. For example, if
swings that the ‘rule of 90’ is violated, the cost of
the aggregate costs 50% more, instead of $0.30/m
binder is increased by about 30%, if SHRP
it will cost $0.45/m , and the price of the chip
recommendations are followed. For a 10 cm thick
seal will rise by $0.15/m . If the energy, en-
new road, the cost of ordinary asphalt is $2/m
vironmental and durability beneﬁts over the life-
and higher grade asphalt costs $2.60/m . Instead
time of the resurfacing exceed $0.15/m , the
of buying the higher grade binder, one could
cooler pavement pays for itself. Again, this de-
apply a chip seal, which costs about $0.60/m . pends on local circumstances: the climate and
Chip seals comprise a binder onto which aggre- smog conditions vs. the cost of light-colored
gate is pressed. The aggregate is visible from the aggregate. For Los Angeles, we have estimated
outset, and, if it is reﬂective, the pavement stays that energy and environmental savings alone are
cooler. It might be sufﬁciently cool that it is
about $0.36/m (present value over the lifetime
unnecessary to use the higher grade binder. For of 5 years for a resurfacing), and, thus, one could
example, the data of Fig. 6 show that a 0.25 afford to pay twice the usual price for aggregate
increase in albedo can reduce the pavement and still have no net increase in cost. Lifetime
temperature by 108C. This suggests that the beneﬁts would also accrue in addition to energy
maximum temperature speciﬁcation for the pave- and smog beneﬁts.
ments might be reduced by 108C. A lower grade
of binder might then be acceptable. The reduced
cost of the binder cancels the cost of the chip seal, 7. CONCLUSIONS
and one enjoys the cooling beneﬁt at no extra Cool surfaces (cool roofs and cool pavements)
cost. and urban trees can have a substantial effect on
Thus, for thick pavements, the energy and urban air temperature and, hence, can reduce
smog savings may not pay for whiter roads. cooling-energy use and smog. We estimate that
However, if the lighter-colored road leads to about 20% of the national cooling demand can be
substantially longer lifetime, the initial higher cost avoided through a large-scale implementation of
may be offset by lifetime savings. An example of heat-island mitigation measures. This amounts to
this is to be seen when a higher grade binder is 40 TWh/year savings, worth over $4B per year
replaceable by a white surface. This must be by 2015, in cooling-electricity savings alone.
evaluated according to the demands on the road Once the beneﬁts of smog reduction are ac-
and the climate. counted for, the total savings could add up to over
Thin pavements. At some times in its $10B per year.
life, a pavement needs to be maintained, i.e., Achieving these potential savings is conditional
resurfaced. This offers an opportunity to get on receiving the necessary federal, state, and local
cooler pavements economically. Good mainte- community support. Scattered programs for plant-
nance practice calls for resurfacing a new road ing trees and increasing surface albedo already
after about 10 years (Dunn, 1996) and the lifetime exist, but to start an effective and comprehensive
of resurfacing is only about 5 years. Hence, campaign would require an aggressive agenda.We
within 10 years, all the asphalt concrete surfaces have started to collaborate with the American
in a city can be made light-colored. As part of this Society for Testing of Materials (ASTM) and the
regular maintenance, any additional cost of the industry, to create test procedures, ratings, and
whiter material will be minimized. Note also that labels for cool materials. We have also initiated
because the lifetime of the resurfacing is only plans to incorporate cool roofs and trees into the
about 5 years, the present value of the savings is Building Energy Performance Standards of AS-
ﬁve-times greater than the annual savings. Thus, HRAE (American Society of Heating Refrigera-
for LA, the present value is about $0.36/m tion, and Airconditioning Engineers), California
($0.03/ft ). Can a pavement be resurfaced with a Title 24, and the California South Coast’s Air
light color at an added cost less than this saving? Quality Management Plans. We also plan to
For resurfacing, there are the options of a black demonstrate savings in selected ‘Cool Com-
topping, such as a slurry seal, or a lighter-colored munities’, including federal facilities, particularly
Cool surfaces and shade trees to reduce energy use and improve air quality in urban areas 309
Bretz S., Akbari H. and Rosenfeld A. (1997) Practical issues
military bases. A related effort involves expand- for using high-albedo materials to mitigate urban heat
ing the Los Angeles Basin’s Regional Clean Air islands. Atmospheric Environment 32(1), 95–101.
Bretz S. and Akbari H. (1997) Long-term performance of
Incentive Market (RECLAIM) NO -credit trading
high-albedo roof coatings. Energy and Buildings Special
market to include air temperature reduction by Issue on Urban Heat Islands and Cool Communities 25(2),
cool surfaces. The South Coast Air Quality Man- 159–167.
Bretz S. and Akbari H. (1994) Durability of high-albedo roof
agement District and the EPA now recognize that coatings. In Proceedings of the ACEEE 1994 Summer Study
air temperature is as much a cause of smog as on Energy Efﬁciency in Buildings, Vol. 9, p. 65.
NO or volatile organic compounds, so that cool Cominsky R. J., Huber G. A., Kennedy T. W. and Anderson
M. (1994). The Superpave Mix Design Manual for New
surfaces and shade trees should be monetized on Construction and Overlays
, National Re-
RECLAIM along with NO . Finally, EPA is
search Council, Washington, DC.
DeWalle D. R., Heisler G. M. and Jacobs R. E. (1983) Forest
considering mechanisms that would allow inclu- home sites inﬂuence heating and cooling energy. Journal of
sion of cool surface and trees in State Im- Forestry 81(2), 84–87.
plementation Plans (SIPs) for ozone compliance. Dunn B. H. (1996) What you need to know about slurry seal.
Better Roads March, 21–25.
EPA (1986). Guideline on Air Quality Models
Environmental Protection Agency EPA-450/2-78-027R.
Acknowledgements—This work was supported by the Assis- Gartland L. M., Konopacki S. J., Akbari H. (1996) Modeling
tant Secretary for Energy Efﬁciency and Renewable Energy, the effects of reﬂective rooﬁng. In Proceedings of the
Ofﬁce of Building Technologies of the U.S. Department of ACEEE Summer Study on Energy Efﬁciency in Buildings,
Energy, and the U.S. Environmental Protection Agency under Vol. 4, p. 117.
Contract No. DE-AC0376SF00098. Gery M. W.,Whitten G. Z. and Kills J. P. (1988) Development
and Testing of the CBM-IV for Urban Abd Regional
Modeling. Report EPA-600/3/ 88-012. U.S. EPA, Research
Triangle, North Carolina. ANL No. 058719, Argonne
National Laboratory, Argonne, IL.
REFERENCES Goodridge, J. (1989). Air temperature trends in California,
1916 to 1987. J. Goodridge, 31 Rondo Ct., Chico CA
Akbari H., Gartland L. and Konopacki S. (1998) Measured 95928.
energy savings of light-colored roofs: results from three Goodridge, J., 1987. Population and temperature trends in
California demonstration sites. In Proceedings of the
California. Proceedings of the Paciﬁc Climate Workshop,
ACEEE Summer Study on Energy Efﬁciency in Buildings Paciﬁc Grove CA, March 22–26.
Vol. 3, p. 1. Hall J. V., Winer A. M., Kleinman M. T., Lurmann F. M.,
Akbari, H., Davis, S., Dorsano, S., Huang, J. and Winnett, S. Brajer V. and Colome S. D. (1992) Valuing the health
(1992). Cooling Our Communities
A Guidebook on Tree beneﬁts of clean air. Science 255, 812–817.
Planting Light-Colored Surfacing. U.S. Environmental Heisler G. M. (1989). In Effects of Trees On Wind and Solar
Protection Agency, Ofﬁce of Policy Analysis, Climate Radiation in Residential Neighborhoods
Final Report On
Change Division Site Design and Microclimate Research
Akbari H., Rosenfeld A. and Taha H. (1990) Summer heat Argonne National Laboratory, Argonne, IL.
islands, urban trees, and white surfaces. ASHRAE Trans
Hildebrandt E., Bos W. and Moore R. (1998) Assessing the
96(1). American Society of Heating, Refrigeration, and Air impacts of white roofs on building cooling loads: a com-
Conditioning Engineers, Atlanta, Georgia. parison of results from three different methods. ASHRAE
Akbari H., Bretz S., Taha H., Kurn D. and Hanford J. (1997) Trans
104(1). Proceedings of the American Society of
Peak power and cooling energy savings of high-albedo Heating, Refrigeration, and Air Conditioning Engineers,
roofs. Energy and Buildings-Special Issue on Urban Heat Atlanta, Georgia (January).
Islands and Cool Communities 25(2), 117–126. Huang Y. J., Akbari H. and Taha H. (1990) The wind-
Akbari H., Bretz S. E., Hanford J.W., Kurn D. M., Fishman B. shielding and shading effects of trees on residential heating
L., Taha H. G. and Bos W. (1993) Monitoring Peak Power and cooling requirements. ASHRAE Trans
and Cooling-energy Savings of Shade Trees and White ings of the American Society of Heating, Refrigeration, and
Surfaces in the Sacramento Municipal Utility District Air conditioning Engineers, Atlanta, Georgia (February).
and Konopacki, S., Akbari, H., Gartland, L. and Rainer, L. (1998).
Results, Lawrence Berkeley Laboratory Report LBL-34411, Demonstration of Energy Savings of Cool Roofs. Lawrence
Berkeley, CA. Berkeley National Laboratory Report LBNL-40673, Ber-
Akbari H. and Taha H. (1992) The Impact of trees and white keley, CA
surfaces on residential heating and cooling energy use in Konopacki, S., Akbari, H., Gabersek, S., Pomerantz, M. and
four Canadian cities. Energy
the International Journal Gartland, L. (1997) Cooling Energy Saving Potentials of
17(2), 141–149. Light-Colored Roofs for Residential and Commercial Build-
Asphalt-Institute (1989). The Asphalt Handbook, The Asphalt ings in 11 U.S. Metropolitan Areas. Lawrence Berkeley
Institute, Lexington, KT. National Laboratory Report LBL-39433, Berkeley, CA
Bally, M., 1998. Private communication on the prices for McPherson, E.G., Nowak, D.J. and Rowntree, R.A. (1994).
Superpave grades of asphalt charged by Koch Performance Chicago’s urban forest ecosystem: results of the Chicago
Asphalt Co. in Nevada. Urban Forest Climate Project, Forest Service, U.S. Dept. of
Benjamin M. T., Sudol M., Bloch L. and Winer A. M. (1996) Agriculture. NE-186.
Low-emitting urban forests: A taxonomic methodology for Means R. S. (1996). Site and Landscape Costs, R.S. Means
assigning isoprene and monoterpene emission rates. Atmos-Company, Inc, Kingston, MA.
pheric Environment 30(9), 1437–1452. Milford J. B., Armistead G. R. and McRae G. J. (1989) A new
Berdahl P. and Bretz S. (1997) Preliminary survey of the solar approach to photochemical pollution control: implications of
reﬂectance of cool rooﬁng materials. Energy and Buildings spatial patterns in pollutant responses to reduction in
Special Issue on Urban Heat Islands and Cool Communities nitrogen oxides and reactive organic emissions. Environ-
25(2), 149–158. mental Science and Technology 23, 1290–1301.
310 H. Akbari et al.
Parker D. S., Sherwin J. R. and Sonne J. K. (1998) Measured and labels for paints. In Proceedings of the ACEEE
performance of reﬂective rooﬁng systems in Florida com- Summer Study on Energy Efﬁciency in Buildings, Vol. 9, p.
mercial buildings. ASHRAE Trans
104(1). American Socie- 141.
ty of Heating, Refrigeration, and Air Conditioning En- Taha H. (1996) Modeling the impacts of increased urban
gineers, Atlanta, Georgia (January). vegetation on the ozone air quality in the South Coast Air
Basin. Atmospheric Environment 30(20), 3423–3430.
Parker D. S., Barkaszi S. F., Jr., Chandra S. and Beal D. J. Taha H. (1997) Modeling the impacts of large-scale albedo
(1995) Measured cooling energy savings from reﬂective changes on ozone air quality in the South Coast Air Basin.
rooﬁng systems in Florida: ﬁeld and laboratory research Atmospheric Environment 31(11), 1667–1676.
results, Proceeding of the Thermal Performance of the Taha H., Douglas S. and Haney J. (1994) The UAM sensitivi-
Exterior Envelopes of Buildings VI, December 4–8, 1995, ty analysis: the August 26–28 1987 oxidant episode. In
Clearwater, FL. Analysis of Energy Efﬁciency and Air Quality in the South
Parker J. H. (1981). Use of Landscaping For Energy Con-Coast Air Basin-Phase II, Taha H. et al. (Ed.), Lawrence
servation, Department of Physical Sciences, Florida Interna- Berkeley Laboratory Report LBL-35728, Berkeley, CA,
tional University, Miami, FL. Chapter 1.
Pielke R. (1974) A three-dimensional numerical model of the Taha H., Douglas S., Haney J., Winer A., Benjamin M., Hall
sea breeze over South Florida. Monthly Weather Review D., Hall J., Liu X. and Fishman B. (1995) Modeling the
102, 115–139. Ozone Air Quality Impacts of Increased Albedo and Urban
Pomerantz M., Akbari H. Chen A., Taha H. and Rosenfeld A. Forest in the South Coast Air Basin, Lawrence Berkeley
H. (1997) Paving Materials for Heat Island Mitigation, Laboratory Report LBL-37316, Berkeley, CA.
Lawrence Berkeley National Laboratory Report LBNL- Taha H., Konopacki S. and Gabersek S. (1996) Modeling the
38074. Berkeley, CA. Meteorological and Energy Effects of Urban Heat Islands
Rosenfeld A. H., Romm J. J., Akbari H. and Pomerantz M. and their Mitigation
-Region Study, Lawrence Ber-
(1998) Cool communities: strategies for heat islands mitiga- keley Laboratory Report LBL-38667, Berkeley, CA.
tion and smog reduction. Energy and Buildings 28, 51–62. Yoder E. J. and Witzak M. W. (1975 ). Principles of Pavement
Rosenfeld A., Akbari H., Taha H. and Bretz S. (1992) Design, Wiley, New York.
Implementation of light-colored surfaces: proﬁts for utilities