Content uploaded by Haochen Tan
Author content
All content in this area was uploaded by Haochen Tan on Nov 29, 2022
Content may be subject to copyright.
Impact of different roofing mitigation strategies on near-surface temperature
and energy consumption over the Chicago metropolitan area during a
heatwave event
Haochen Tan
a,
⁎,Rao Kotamarthi
a
, Jiali Wang
a
, Yun Qian
b
, T.C. Chakraborty
b
a
Environmental Science Division (EVS), Argonne National Laboratory, Lemont, IL, United States
b
Atmospheric Sciences and Global Change Division, Pacific Northwest National Laboratory, Richland, WA, United States
HIGHLIGHTS GRAPHICAL ABSTRACT
•The use of solar photovoltaic panels has
surged in the citie s recently.
•Urban atmosphere model compares sce-
narios for installing cool roofs, green
roofs, and solar panel roofs during
heatwave event.
•Best cooling energy saving when solar
panel roofs are applied.
ABSTRACTARTICLE INFO
Editor: Shuqing Zhao
Keywords:
Heatwave
Rooftops
Cooling energy demand
Urban climate modeling
This study examined the impact of cool roofs, green roofs, and solar panel roofs on near-surface temperature and
cooling energy demand through regional modeling in the Chicago metropolitan area (CMA). The new parameteriza-
tion of green roofs and solar panel roofs based on model physics has recently been developed, updated, and coupled
to a multilayer building energy model that is fully integrated with the Weather Research and Forecasting model. We
evaluate the model performance against with observation measurements to show that our model is capable of being
a suited tool to simulate the heatwave event. Next, we examine the impact by characterizing the near-surfaceair tem-
peratureand its diurnal cycle fromexperiments with and without the different rooftops.We also estimate the impactof
the rooftop on the urban island intensity (UHII), surface heat flux, and the boundary layer. Finally, we measure the im-
pact of the different rooftops on citywide air-conditioning consumption. Results show that the deployment of the cool
roof can reduce the near-surface temperature most over urban areas, followed by green roof and solar panel roof. The
cool roof experim ent was the only one where the near-surface temperature trended down as the urban fraction
increased, indicating the cool roof is the most effective mitigation strategy among these three rooftop options. For
cooling energy consumption, it can be reduced by 16.6 %, 14.0 %, and 7.6 %, when cool roofs, green roofs, and
solar panel roofs are deployed, respectively. Although solar panel roofs show the smallest reduction in energy
consumption, if we assume that all electricity production can be applied to cooling demand, we can expect almost a
savings of almost half (46.7 %) on cooling energy demand.
Science of the Total Environment xxx (xxxx) xxx
⁎Corresponding author at: Argonne National Laboratory, 9700 S. Cass Ave., Lemont, IL 60439, United States.
E-mail address: htan@anl.gov (H. Tan).
STOTEN-160508; No of Pages 14
http://dx.doi.org/10.1016/j.scitotenv.2022.160508
Received 9 September 2022; Received in revised form 31 October 2022; Accepted 22 November 2022
Available online xxxx
0048-9697/Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Contents lists available at ScienceDirect
Science of the Total Environment
journal homepage: www.elsevier.com/locate/scitotenv
Please cite this article as: H. Tan, R. Kotamarthi, J. Wang, et al., Impact of different roofing mitigation strategies on near-surface temperature an..., Sci-
ence of the Total Environment, http://dx.doi.org/10.1016/j.scitotenv.2022.160508
1. Introduction
Deployment of cool or green roofing technology on a broad scale can
mitigateurban heat and reduce energy consumption and has been proposed
widely for these purposes (Akbari et al., 2009;Salamanca et al., 2012a;Li
et al., 2014;Santamouris, 2014;Georgescu et al., 2014;Zhang et al.,
2017;Tan et al., 2019;Zhang et al., 2019;Zonato et al., 2021). Both cool
and green roofs can reduce heat storage in buildings (thus cooling them),
but their mechanisms differ. Because cool roofs increase albedo (or surface
reflectivity), they absorb less incoming shortwave radiation than conven-
tional roofs, leading less heat into the urban canopy and eventually reduc-
ing surface temperature through the mixing of air. Consequently, cool roofs
limit the transmission of heat into urban interiors and the broader environ-
ment, reducing air-conditioning consumption and near-surface tempera-
ture, respectively. Green roofs, on the other hand, can convert available
energy more efficiently than conventional roofs into latent heat flux
through vegetation by increasing evapotranspiration on the rooftop and
lowering sensible heat flux transfer into the urban surface layer.
The immediate advantages of large-scale implementation of cool and
green roofs in urban areas have been demonstrated at a building or neigh-
borhood scale (Wong and Chen, 2005;Jaffal et al., 2012;Sun et al., 2013,
2014;Kong et al., 2016). However, because of the significant influence of
surface heterogeneity on the local microclimate, the effects of mitigation
approaches applied at these scales may not easily translate to city-scale ben-
efits (Bou-Zeid et al., 2004, 2007;Li et al., 2014). Furthermore, upscaling
from the buildings to the city may not capture the impacts of individual
building rooftops on outside urban air temperatures, the exchanges of air
temperature/moisture, and surface energy balances (Li et al., 2014).
Numerical weather prediction (NWP) models have been used
frequently to quantify the impacts of green/cool roofs at the city scale
(Synnefa et al., 2008;Millstein and Menon, 2011;Georgescu et al., 2013;
Li et al., 2014;Li and Norford, 2016;Sun et al., 2016;Zhang et al., 2017;
Yang et al., 2016;Morini et al., 2017;Zonato et al., 2021). Millstein and
Menon (2011) showed that nationwide cool roof deployment over the
United States would lower summertime air temperature by 0.11–0.53 °C
based on 12 simulated summer periods using a model with a 25-km grid
spacing. Georgescu et al. (2013) demonstrated that the regional warming
over Arizona caused by urbanization could be moderated by 50 % by
using cool roofs. For the Baltimore-Washington region, it has been shown
that surface air temperature could be lowered by 3.5 °C by green roofs
and 2 °C by cool roofs (Li et al., 2014). Over the Yangtze River Delta in
China, researchers found that green and cool roofs could lower the daily
mean surface air temperature by 1 °C and 3.8 °C, respectively (Zhang
et al., 2017).
However, cool roofs and green roofs are not the only roofing options
that can mitigate heat and reduce electricity consumption. Recently, the
use of solar photovoltaic panels has surged in the cities. Dominguez et al.
(2011) showed that in San Diego, California, solar photovoltaic panels
that partially cover building rooftops can reduce not only greenhouse gas
emissions but also the annual cooling load. Salamanca et al. (2016) found
that solar photovoltaic panels could lower the 2-m temperature over Phoe-
nix and Tucson, Arizona. Ma et al. (2017) found that solar photovoltaic
panels might reduce summer maximum temperatures in Sydney. In a
study that corroborates these findings, Taha (2013) demonstrated that a
vast solar panel deployment acrossthe Los Angeles area hasa cooling effect
of up to 0.2 °C. Conversely, the study by Zonato et al. (2021) exhibits this
differently: solar panels can induce a temperature increase during the
daytime due to less efficient heat release and more sensible heat flux
together from solar panels and the roof.
Earlier studies have presented comprehensive effects of various roofing
technologies, including cool roofs, green roofs, and solar panel roofs, on air
temperature. However, some of these studies focus only on the building or
neighborhood scale. Also, among studies that use NWP models at the urban
scale, some employ only single-layer urban parameterization instead of
multi-layer urban parameterization, even though the multi-layer option
can simulate the three-dimensional urban structure. To fully examine the
impact of different roofing mitigation strategies on urban climate and
energy demand, we use a high-resolution regional scale modeling system
(500 m grid spacing) fully coupled with a building energy model. The spa-
tial resolution of the World Urban Database and Access Portal Tools
(WUDAPT, Ching et al., 2018) is 100 m, which it has rendered to NWP
model resolution (500m) in this study. Our goal is to evaluate and compare
the regional impacts of the extensive deployment of cool roofs, green roofs,
and solar photovoltaic panel roofs on near-surface temperature and air-
conditioning (AC) consumption in the Chicago region. To the best of our
knowledge, little research has been done to evaluate these three roof
mitigation strategies collectively over Chicago.
In this paper, we introduce the mechanism of the three roof technolo-
gies and thus assess the impacts of the cool roof, green roof, and solar
panel roofs technologies on near-surface temperature and energy consump-
tion from air-conditioning by using the building model used Building Effect
Parameterization coupled with the Building Energy Model (BEP + BEM,
Martilli et al., 2002;Salamanca and Martilli, 2010) in Weather Research
and Forecasting (WRF) regional climate model (version 4.3.1; Skamarock
et al., 2021).We first evaluate the model performance against with observa-
tion measurements to show that our model is capable of being a suited tool
to simulate extreme weather conditions. Next, we examine the impact by
characterizing the near-surface air temperature and its diurnal cycle from
experiments with and without the rooftop. We also estimate the impact of
the rooftop on the urban island intensity (UHII) and surface heat flux, and
finally, we measure the impact of the different rooftops on citywide air-
conditioning consumption. We characterize these factors for different
roofs in the Chicago metropolitan area (CMA). The 1995 heatwave in the
CMA resulted in the highest recorded mortality rates during a heat wave
in North America (Livezey and Tinker, 1996). In recent decades, an increas-
ing warming trend has been documented in the region, resulting in
increased human health hazards and energy consumption (Chen et al.,
2022). The regional climate of CMA is moderated by its adjacency to
Lake Michigan and the natural lake breeze occasionally mitigates UHI
(Davies et al., 2007;Harris and Kotamarthi, 2005). However, the lake
breeze is not an effective solution most days and for all the CMA and it is
important to understand whether the broad deployment of rooftop mitiga-
tion strategies over CMA would be effective forUHI mitigation in theface of
the increasing warming trends. Methodology and simulation designs are
outlined in Section 2. The results are in Section 3, and the summary and
conclusions are in Section 4.
2. Methodology
2.1. Modeling system and land use
We ran a regional climate model combined with a multi-layer building
system over CMA. The model for the climate was the non-hydrostatic and
fully compressible Weather Research and Forecasting (WRF) regional
climate model (version 4.3.1; Skamarock et al., 2021). The building
model used Building Effect Parameterization coupled with the Building
Energy Model (BEP + BEM, Martilli et al., 2002;Salamanca and Martilli,
2010) was used to simulate the diurnal variation of near-surface tempera-
ture and cooling energy consumption over the city. The BEP model was de-
veloped and validated offline by Salamanca et al. (2010) and implemented
in Salamanca et al. (2010). The BEP + BEM system calculates the surface
momentum, heat exchanges, humidity, and turbulent kinetic energy fluxes
to the atmospheric dynamics governing equations under atmospheric
conditions at the WRF bottom level. Hence, the coupling of the BEP +
BEM system with WRF is accomplished. Compared to Single-Layer Urban
Canopy Model (SLUCM, Kusaka et al., 2001), the BEP + BEM model
takes into account 1) the three-dimensional urban structure in a model
grid using multiple layers and the origins and sinks of momentum and
heat in the vertical layers within the urban canopy layer, 2) the impact of
horizontal and vertical surfaces (such as roads and walls) on the momen-
tum, turbulent kinetic energy, and potential temperature, 3) the shading,
reflection and blocking effect of horizontal and vertical surfaces on net
H. Tan et al. Science of the Total Environment xxx (xxxx) xxx
2
solar radiation within the urban canopy layer, 4) exchange of heat between
building walls, rooftops, and floors, 5) heat discharged by people and
domestic electrical appliance and 6) air-conditioning cooling, heating,
and ventilation (Salamanca and Martilli, 2010;Ribeiro et al., 2020;
Ricard et al., 2021).
Quantifying the physical processes within an urban region requires an
accurate description of advanced urban geometry. Thus, for this study, an
integrated spatial model of local climate zones (LCZ) was created,a concept
defined and developed by Stewart and Oke (2012). The new version of
WRF-Urban and BEP + BEM urban multi-layer parameterization can inte-
grate 11 urban classifications from WUDAPT. This method has already
been implemented in many studies (Brousse et al., 2016;Zonato et al.,
2020;Hammerberg et al., 2018;McRae et al., 2020;Patel et al., 2020).
The LCZ is the level 0 product produced by WUDAPT and incorporates
different land use categories with comparable long-term meteorological
features. Foley (2015) created the LCZ database for the CMA region,
which categorizes areas into 11 classifications instead of the traditional
three urban categories. Another update in this WRF-Urban is that a new
buildings drag coefficient that induced by buildings for mean wind speed
and turbulent kinetic energy is applied and its improvement to the model
have shown by Gutiérrez et al. (2015) and Santiago and Martilli (2010).
2.2. Numerical modeling of cool roofs, green roofs, and solar panel roofs
Fig. 1 shows the basic concept of a grid cell with built impervious
fractions and different types of roofs for urban buildings. The impervious
part of the urban canopy includes buildings, roads, and pavements. The
surface energy balance for the conventional rooftop is presented by
SWNet þLWNet þAH ¼LH þSH þGð1Þ
where SW
Net
is the net shortwave radiation at the surface, LW
Net
is the net
longwave radiation at the surface, AH is the anthropogenic heat flux, LH
and SH are the latent and sensible heat flux, and G is the ground heat
flux. For BEP + BEM in WRF, the anthropogenic heat flux represented is
from air-conditioning, heat exchanges between the building interior and
exterior air, and the heat released by equipment and people within the
buildings. This is a limitation of BEP + BEM: BEP + BEM treats the factors
above as the only anthropogenic heat sources, where it does not include
anthropogenic heat from traffic or industry. For a traditional roof, all net
radiation is transferred into sensible heat flux and heat flowing into the
building, raising the skin temperature and near-surface temperature. Cool
roofs can lower the net radiation at the roof by reflecting the shortwave
radiation from the reflective roofing material.
The green roof parameterization has been established according to de
Munck et al. (2013) and it was recently updated by Zonato et al. (2021).
The parameterization for green roofs computes energy and water supply,
estimates net radiation, water contribution from precipitation and irriga-
tion, vegetation evapotranspiration, heat transmission, and energy and
moisture distribution through the soil. The green roof scheme in this WRF
version incorporates ten layers including five levels of organic matter
substrate where the plant grows, one layer of drainage, and four levels of
the insulation layer, with a total of ~0.3 m. A complete description of
hydrology and thermodynamics for green roofs can be found in Zonato
et al. (2021).
The parameterization of solar photovoltaic panel roofs in the WRF
model was previously developed by Masson et al. (2014) and the new
parameterization has been developed and tested by the developers in the
latest version of WRF model (Zonato et al., 2021). The solar panels in this
parameterization are assumed to be parallel and unattached from the roof
so that not only the profiles of individual buildings can remain simple but
also the shading effect of solar panels decreases the surface temperature
of the roof from the shading effect The temperature over solar panels roof
and its associated heat fluxes are determined by the derivative equation
below with all terms in W m
−2
:
Cmodule
∂TPV
∂t¼1−αPV
ðÞSW↓
sky þεU
PVLW↓
sky−LW↑
PV
þLW↑↓
roof−PV−EPV −SH↑−SH↓þ1−VFðÞ
1−αPV
ðÞSWDIFF þLW↓
sky
hi ð2Þ
where C
module
is the equivalent heat capacity per unit area, αis albedo. The
terms on the right-hand side are net shortwave radiation gained by the
upward surface of the solar panel; the incoming longwave radiation at the
upper surface of the solar panel where the ε
PV
U
is the emissivity of the
glass face; upward longwave radiation emitted by solar panel; longwave
radiation exchanged between the monocrystalline silicon downward face
of the solar panel and the upward of the roof; energy production by solar
panel; upward sensible heat flux from the solar panel; downward sensible
heat flux from solar panel; diffuse shortwave and longwave isotropic radia-
tions reaching the downward solar panel surface where VF is the view
factor between downward face of the solar panel and the roof. More expla-
nation of this equation can be found in Zonato et al. (2021).
Fig. 1. Schematic description of urbangrid cell applied inWRF for (left to right) conventional roof, cool roof, greenroof, and solar panel roof modeling. H
a
is the height of first
level in the atmospheric model,H
r
is the heightof the building rooftop,and H
c
is the canopy height. SH and LH represent sensible and latent heat flux, respectively. T
a
signifies
the temperature at the beginning level of the atmospheric model. T
c
is the temperature at canopy level. T
g
is the surface temperature. T
wall
is the temperature from the
building wall. SW and LW are shortwave and longwave radiation, respectively, and G
r
is the ground heat flux into the building. E
pv
characterizes the electricity generated
by solar panels. The blue box on the side on the building represents the AC unit. The length of each arrow does not represent the magnitude of each variable. The
detailed description of parameterization is in Section 2.2.
H. Tan et al. Science of the Total Environment xxx (xxxx) xxx
3
The configuration of solar panel roof parameterization is defined as
“PV_FRAC_ROOF”in the URBPARAM_LCZ.TBL to arrange the coverage
fraction over buildings. Compared to the parameterization from Masson
et al. (2014) and Salamanca et al. (2016), the temperature on solar panels
roof in the new parameterization is solved numerically from the derivative
equation instead of parameterizing the temperature through its depen-
dence on shortwave solar radiation, contributed by all the associated com-
ponents (Zonato et al., 2021). Other heat fluxwill be updated and proceeds
to other urban sections after the solar panel roof temperature is computed.
In our solar panel experiment, we assume the roofs are entirely covered
with solar panels (i.e., the fraction of solar panel equal 1).
2.3. Model set-up and experiment design
This study targets CMA and its neighboring rural areas, centered on
41°52′54″N, 87°37′23″W. Three two-way nested domains were used with
406 × 313 grid points in domain 1, 280 × 382 in domain 2, and 292 ×
583 in domain 3, with grid spacing of 4.5, 1.5, and 0.5 km in WRF 4.3.1, re-
spectively (Fig. 2a). The outermost parent domain includes the entire Great
Lakes, and the innermost domain encloses CMA. The LCZ classifications are
integrated in the WRF model by updating the surface parameters for every
grid in the innermost domain. As such, the initial MODIS land use categories
have been added an extra 11 urban categories to allow WRF to use the spe-
cialized categories in the LCZ model we have chosen (Fig. 2b). The details
including the workflow about how to ingest LCZ information into WRF
can be found in Demuzere et al. (2020, 2022). Non-urban land covers are
based on the Moderate Resolution Imaging Spectroradiometer (MODIS)
21-class product (Kumar et al., 2014). For the length of the simulation
(6 days), the size of the outermost domain was sufficient to capture synoptic
phenomena throughout the domain. We use the 51 eta levels for vertical
grid-spacing as in Salamanca et al. (2012a,b): these range from surface to
100 hPa, with 38 levels located below 2 km and with the model top over
about 20 km. The integration time step was 20 s for the outermost domain,
and the simulation outputs were collected every 15 min for the innermost
domain. The initial and boundary conditions were taken from the National
Centers for Environmental Prediction Final Analysis data (NCEP-FNL, 1° ×
1°, 6-hourly). In this study, we used the unified Noah land-surface model
for the land-surface scheme (Chen and Dudhia, 2001;Ek et al., 2003), the
Mellor-Yamada-Janjic scheme for the planetary boundary layer (Janjic,
2002), the WRF Single Moment 6 for microphysics (Hong and Lim, 2006),
the Dudhia scheme for shortwave radiation (Dudhia, 1989), the Rapid
Radiative Transfer Model for longwave radiation (Mlawer et al., 1997),
and the Monin-Obukhov-Janjic approach (Janjic, 2002) for the surface-
layer scheme. No cumulus parameterization was used.
Four experiments were conducted, all including the same 6-day, clear-
sky period from August 21 (0000 LT) to August 27 (0000 LT), 2021, to
estimate the regional effects of extensive deployment of different roofs on
near-surface temperature and AC consumption. A heatwave event occurred
on August 24, 2021, with the temperature reaching a high of 35 °C (95 °F)
in the Chicago area. The control simulation (CTL, Table 1)wasrunusinga
roofs albedo of 0.2 as the baseline case, with no green roofs/solar panels.
The cool roof simulation (CR, Table 1) emulated the impact of cool roofs
by settingthe urban roof albedo to 0.8. Similarly, the green roof simulation
(GR, Table 1) addressed the impact of the green roofs by setting the green
roof fraction to 1 for each urban roof. Finally, the solar panel roof experi-
ment (SPR, Table 1) set each urban roof's solar panel fraction equal to 1.
The albedo, conversion efficiency, and emissivity for solar panel roofs
were set to 0.2, 0.19, and 0.90, respectively, which are standard numbers
for the current solar panel equipment. Based on the study by Demuzere
et al. (2020), we altered the default urban parameters used in the urban
parameterization, namely, urban fraction, road and roof width, thermal
properties, heat capacity, and emissivity. The values applied for these sim-
ulations are summarized in Table 2. The thermal properties of different
roofs are demonstrated in Table 3, taken from Sun et al. (2013),Li et al.
(2014),andZo nato et al. (2021).Table 4 summarized the urban parameters
that were applied in this study. We excluded the coverage of cool roofs,
green roofs, and solar panel roofs on three LCZ land use categories: Sparsely
Built (LCZ 9), Heavy Industrial (LCZ 10), and Rock and Paved (LCZ E), as
these three categories are not ideal for roof deployment.
2.4. Observational data
To evaluate the model performance for near-surface meteorologies,
such as 2-m temperature and 10-m wind speed, we compared observations
from six weather stations (Table 5) over CMA against the simulated values.
Fig. 2. (a) WRF domain configuration. The outer domain (1), themiddle domain (2), and the inner domain (3) have a resolution of4.5 km, 1.5 km, and 500 m, respectively.
(b) Urban land-use categories (shading) after the incorporation of local climate zones (LCZs) from WUDAPT data over domain 3. (c) Same as in (b) but for Chicago
Metropolitan Area. The domain in (c) is the area used the analysis.
Table 1
Summary of experiments performed.
Experiment Roof type
Control (CTL) Conventional roof with albedo equal to 0.2
Cool roof (CR) Roof with albedo equal to 0.8
Green roof (GR) Green roof fraction equal to 1
Solar panel roof (SPR) Solar panel roof fraction equal to 1
H. Tan et al. Science of the Total Environment xxx (xxxx) xxx
4
Four of the stations are considered urban and two are rural. It is important
to highlight that, even though the experiment has a high resolution (500 m
for the innermost domain) compared to earlier research, more spatial
variability should be expected in the observations compared to the model
because a 500 m × 500 m grid cell can still include various land-use
types; hence, misclassifications could exist between a grid cell's land-use
type and a station's footprint (Sharma et al., 2016). Daytime is defined as
1100 to 1700 local standard time (LST) and nighttime as 0000 to 0600 LST.
2.5. Quantifying urban heat island intensity (UHII)
There are various methods to calculate the UHII (Liao et al., 2018;Li
et al., 2019;Chen et al., 2022). Here we use the method of Li et al.
(2019), where the relationship between temperatureand the urban fraction
(FRC_URB) is a linear equation given by
T¼FRC URB UHII þTvegetation ð3Þ
where the urban fraction becomes the independent variable. The slope of
this equation is the UHII, and the intercept is the baseline temperature
over a completely pervious surface, which in this region is primarily crop-
lands. We exclude areas that have open water bodies in this calculation.
The FRC_URB is acquired from MODIS and LCZ in WUDAPT; the value
reflects the degree of urbanization and the portion of impervious areas.
2.6. Assessing heat index (HI)
The urban heat island effect is linked with thermal comfort and hence
influences people's health conditions (Mavrogianni et al., 2011). In our
study, we examine the heat index (HI) by using a multiple regression anal-
ysis that was first used by Steadman (1979) and again in many other recent
studies (Yip et al., 2008;Mohan et al., 2014). The Steadman's HI is widely
used by U.S. National Weather Service as an operational metric of thermal
comfort that includes temperature and relative humidity together. It
describes the perception of heat that a human body feels under the current
weather. The calculation of HI is as follows (Rothfusz, 1990):
HI ¼−42:379 þ2:04901523 Tþ10:14333127 RH
−0:22475541 TRH−0:00683783 T2−0:05481717 RH2
þ0:00122874 T2RH þ0:00085282 TRH2−0:00000199 T2RH2
ð4Þ
where the HI is calculated in °F but has transferred to °C for consistency, T is
the dry-bulb temperature, and RH is the relative humidity.In our study,the
HI was calculated based on 2-m temperature and 2-m relative humidity
from WRF outputs.
3. Results
3.1. Model validation: near-surface temperature and wind
The near-surface temperature above ground is calculated using the
Monin-Obukhov similarity theory (Monin and Obukhov, 1954)as:
T2¼TsþTa−Ts
ðÞ
U
Uð5Þ
where T
s
is the surface temperature, T
a
is the 1st layer modellevel, U⁎is the
friction velocity at 2 m and U is the velocity at the 1st model level. The CTL
simulated 2-m air temperature and 10-m horizontal wind were evaluated
against values from 6 observation stations (Fig. 3) over the entire simula-
tion period (August 21st to 27th, 2021). The results present that the CTL
agreeably simulated the variations of near-surface temperature and 10-m
Table 3
Thermal properties of different roofs. For solar panel roof, the depth of the solar
panel is 6.55 mm. The solar panel height away from the building's top surface is
80 cm. Detail can be found in Zonato et al. (2021).
Cool roof Green roof Solar panel roof
Albedo (−) 0.8 0.25 0.2
Emissivity (−) 0.9 0.9 0.9
Heat capacity (MJ m
−3
K
−1
) 2.0 1.9 5.72
Thermal conductivity (J m
−1
s
−1
K
−1
) 1.0 1.1 1.0
Roof depth (cm) 20 30 0.0655 + 80
Table 4
Urban canopy parameter data associated with LCZ types in WRF.
LCZ Urban
fraction
Vegetation
fraction
Average building
height (m)
Road
width (m)
1. Compact highrise 1 0 35 15
2. Compact midrise 0.95 0.05 16.75 12.7
3. Compact lowrise 0.9 0.1 7.25 5.7
4. Open highres 0.65 0.35 33.5 37.5
5. Open midrise 0.7 0.3 19 33.3
6. Open lowrise 0.65 0.35 6.75 12.4
7. Lightweight lowrise 0.85 0.15 5 2.0
8. Large lowrise 0.85 0.15 8.25 32.5
9. Sparsely built 0.3 0.7 6.25 10.5
10. Heavy industry 0.55 0.45 11 28.5
Table 5
Name, location, and land-use class of observational stations used in this study (from MesoWest website).
Station Station name Lat., Lon. (°N, °W) Land-use classification in WRF
1 E1098 Chicago 41.88183, −87.66333 LCZ 2, compact mid-res
2 KMDW Chicago Midway Airport 41.78417, −87.75528 LCZ 8, large low-res
3 KORD Chicago O'Hare Airport 41.97972, −87.90444 LCZ 6, open low-res
4 KPWK Chicago Wheeling 42.12083, −87.90472 LCZ 6, open low-res
5 DeKalb Taylor Municipal Airport 41.93381, −88.70657 Rural (cropland)
6 Aurora Municipal Airport 41.77132, −88.48147 Rural (cropland)
Table 2
Modified urban parameters that used in this study for 4 main urban categories
(LCZ3, LCZ5, LCZ6, LCZ8), which also include the thermal properties for conven-
tional roof on each category.
LCZ name and designation LCZ 3 LCZ 5 LCZ 6 LCZ 8
Urban fraction (−) 0.9 0.7 0.65 0.85
Road and roof width (m) 5.7 33.3 12.4 32.5
Surface albedo of road/roof/wall (−) 0.2/0.2/0.15
Thermal conductivity of road (J m
−1
s
−1
K
−1
) 1.00 1.25 1.00 1.25
Thermal conductivity of roof (J m
−1
s
−1
K
−1
) 1.25 1.45 1.25 1.25
Thermal conductivity of walls (J m
−1
s
−1
K
−1
) 0.69 0.62 0.60 0.80
Heat capacity of road (J m
−3
K
−1
) 1.63e6 1.50e6 1.47e6 1.8e6
Heat capacity of roof (J m
−3
K
−1
) 1.44e6 1.8e6 1.44e6 1.8e6
Heat capacity of walls (J m
−3
K
−1
) 2.05e6 2.0e6 0.72e6 1.8e6
Emissivity of road (−) 0.95
Emissivity of roof (−) 0.90
Emissivity of walls (−) 0.90
H. Tan et al. Science of the Total Environment xxx (xxxx) xxx
5
horizontal wind (Fig. 3). The mean bias error (MBE) of CTL compared with
6 observational stations is 1.4 °C over the entire simulated period. The sim-
ulated 2-m temperature averaged across 6 observational stations shows a
warm bias at night. These biases were the systematic biases of the model
and were identified by several earlier studies (Lee et al., 2011;Kim et al.,
2013;Chen et al., 2014;Giovannini et al., 2014;Janicke et al., 2017;
Wang et al., 2022). Furthermore, compared to BEP, theair conditioning sys-
tems for indoor cooling that are recognized in the BEP + BEM scheme as
anthropogenic heat release produce more heat to be released into the atmo-
sphere (Ribeiro et al., 2020). As a result, the urbanized regions are warmer
at night than they are in the BEP experiments (Ribeiro et al., 2020). For
wind speed, the CTL captures this variation in wind speed, although the
speed is lower than that in the observation during afternoon hours. The
low bias in wind speed is possibly due to misclassifications that could
exist between a grid cell's land-use type and a station's footprint. Several
of the stations, even those within CMA, are enclosures over relatively flat
terrain, which a multi-layer parameterization cannot capture at the same
scale. Further studies involving high-quality input data with more detailed
land use classification, more accurate information on urban canopy struc-
ture and on urban surface's physical/thermal characteristics; as well as
higher horizontal model resolution are strongly recommended to perform
better in urban simulations. The performance of CTL during the heat
wave event (Aug 24th), as represented by MBE and RMSE compared to
the urban and rural stations, is summarized in Table 6.
3.2. Impacts of roof strategies on near-surface temperature and UHII
Fig. 4a–f shows the impacts of cool roofs, green roofs, and solar panel
roofs deployment on near-surface temperature. All experiments show a
cooling effect during the daytime. For example, the CR reduces the near-
surface temperature by 1.5 °C during the daytime over CMA, followed by
a 1.2 °C decrease for GR and a 0.6 °C decrease for SPR. The pattern is
different at night: the GR has a warming effect at nighttime (Fig. 4e, f,
and g), and for the CR experiment, the near-surface temperature remains
relatively unchanged at night (compared with CTL). The SPR experiment
also shows slight nighttime warming at the south of the CMA. In general,
the roofing mitigation strategies reduce the near-surface temperature
during the daytime but become less effective at night. These findings are
consistent with previous analyses by Scherba et al. (2011) over Chicago,
Georgescu (2015) over California, Yang et al. (2016) over Houston,
Texas, and Zonato et al. (2021) in an idealized case. The nighttime differ-
ence can be explained by the following mechanism: because the net
radiative flux is larger in summer than in winter, the urban infrastructure
can collect more heat flux during the day and release it during the night.
Because the green roof has additional soil layers, it is able to store more
heat during daytime compared to other roofs and release that energy at
night, generating a warming impact. Moreover, when the incoming solar
Fig. 3. Time series of (a) 2-m temperature (in °C) and (b) 10-m wind (in m s
−1
) from observations (black) and CTL simulation (red) averaged across 6 weather stations
between August 21–27th. The x-axis uses the local standard time (LST).
Table 6
Summary of averaged mean bias error (MBE) and root mean square error (RMSE)
between CTL and urban/rural observational stations for 2-m temperature and 10-m
wind during day (0700–1900 LST), night (1900–0700 LST), and total during August
21–27th, 2021.
Time 2-m temperature 10-m wind
MBE RMSE MBE RMSE
Urban stations
1, 2, 3, 4 Total 1.48 2.28 −1.67 2.11
Day 0.24 1.63 −1.08 2.43
Night 1.24 2.79 −0.59 1.74
Rural stations
5, 6 Total 1.54 3.21 −0.64 1.66
Day 0.16 2.40 −0.46 1.60
Night 1.38 3.85 −0.18 1.71
H. Tan et al. Science of the Total Environment xxx (xxxx) xxx
6
radiation is missing at night, the vertical mixing over the urban layer is
inadequate, leading to stable air temperature evolution (Poulos et al.,
2002;Yang et al., 2016). The performance of green roofs and solar panel
roofs has been tested in wintertime by Zonato et al. (2021) and shows
that the green roof and solar panel roof are capable of increasing the
near-surface temperature and hence reducing energy consumption.
In addition, the implementation of roofing technology could affect
different urban land-use categories. Table 7 presents peak heat mitigation
for different urban land-use categories. The peak daily heat island mitiga-
tion is calculated by using the peak 2-m temperature from each roofing
experiment minus the2-m temperature from CTL at the same local standard
time (LST) during the heatwave event (Aug 24th), then calculating the
area-averaged over the black box in Fig. 1. As the urban fraction changes
(from LCZ 6, 8, 3), the mitigation of peak heat island changes too, for all
of the three roofing technologies. Overall, the CR has the best mitigation
effect on near-surface temperature over these three major urban categories,
which is further supported by Zonato et al. (2021).Fig. 5 quantifies the
impact of different rooftops on UHII during daytime. For CTL, the mean
daytime 2-m temperature increases with the increases in FRC_URB, imply-
ing the connection between UHII and urban development level. The CR, on
the other hand, shows that temperature tends todecreaseas the urban frac-
tion increases, indicating that it is an effective rooftop mitigation strategy.
The UHII in CTL is 1.91 °C, followed by −0.27 °C for CR, 0.40 °C for GR,
and0.98°CforSPR.
3.3. Heat index
Fig. 6 presents the 6-day averaged HI difference for different roof exper-
iments. It shows that the CR experiment lowers the HI more than the GR
and SPR experiments do. The reason that the GR condition does not signif-
icantly decrease the HI is that there is higher evaporation during daytime
and a lower horizontal wind speed, which is further explained by Sharma
et al. (2016) over the Chicago area. The CR slightly lowers the HI by
1.4 °C, followed by a 0.8 °C decrease for SPR and a 0.3 °C decrease for
Fig. 4. The 6-day averaged 2-m temperature difference (minus CTL) for (a) Cool Roof, (b) Green Roof, and (c) Solar Panel Roof during daytime (1100–1700 LST).(d–f) show
the difference during nighttime (0000–0600 LST). (g)shows the diurnal cycle of 2-m temperature difference betweencool roof (blue), green roof(green), and solar panel roof
(red) to CTL. Units are °C.
Table 7
Peak daily heat island mitigations (°C) based on 2-m temperature averaged over
Chicago metropolitan area for different land use categories.
LCZ land-use category Cool roof Green roof Solar panel roof
LCZ 6 3.8 2.2 2.1
LCZ 3 3.7 2.4 2.2
LCZ 8 4.4 2.9 2.7
H. Tan et al. Science of the Total Environment xxx (xxxx) xxx
7
GR. Fig. 7 presents the hours of HI difference that have temperatures
>32.2 °C (90 °F) during the heatwave event (August 24, 2021). The results
show that for CR, the HI hours that exceed 32.2 °C are noticeably reduced
over the south of CMA but slightly increased over the northern coastline.
The GR and SPR, however, had less effect on the hours of HI that exceed
32.2 °C over most of CMA. For near-surface relative humidity, all tests indi-
cate an increase over the CMA, with the CR increasing the relative humidity
the most (Fig. 8).
Besides numerical simulations, urban canopy parameters such as sky
view factor (SVF) and frontal area index (FAI) also impact the UHI effect
(Chen and Ng, 2011;He et al., 2019). SVF is defined as the ratio between
the radiation received by a planar surface from the sky to the radiation
released to the total hemispheric region, which is also adimensional. Over
urban areas, the value of SVF reduces with the increase of sky that is
blocked by the building, and hence more longwave radiation is trapped
inside the urban canopy. Consequently, a less SVF is commonly associated
with a more intense UHI. (He et al., 2019). The FAI calculates the frontal
area per unit horizontal area (Burian et al., 2002a,b). The lower FAI leads
to a higher local wind speed and subsequently a stronger mitigation impact
on the UHI effect (Burian et al., 2002a,b;Chen and Ng, 2011). Even though
the impact of these two factors on UHI is beyond the scope of this research,
more comprehensive knowledge can be found in Chen and Ng (2011).
Fig. 5. Relationship betweendaytime 2-m temperature (°C) and the urbanfraction (dimensionless) with linear fitting (red line)across solid black box areain Fi g. 2b.The unit
for urban heat island intensity (UHII) is °C.
Fig. 6. The 6-day averaged heat index difference (minus CTL) for (a) CR, (b) GR, and (c) SPR. Units are °C.
H. Tan et al. Science of the Total Environment xxx (xxxx) xxx
8
3.4. Impacts of roof strategies on boundary layer
To further illustrate the impact of CR, GR, and SPR on the boundary
layer, we present the time series of differences in vertical profiles of relative
humidity, horizontal wind, and temperature by subtracting the results from
the CTL simulation. The reduction of temperature in atmospheric up to
1.4–1.6 km was observed during the heatwave event (Aug 24th) when
roofs are applied (Fig. 9 a–c). The maximum reduction in the atmospheric
temperature occurred when the CR was applied. For horizontal wind, the
decrease in horizontal wind due to the reduced vertical mixing of momen-
tum was detected below 1 km of the atmosphere and the increase above
1 km during the heatwave event (Aug 24th), especially over CR and GR.
After the vertical mixing is decreased, the airflow over higher levels with
stronger wind speed was less penetrated into lower-level air with less
wind speed which is due to less momentum shift from higher levels to
lower levels, so that the wind speed above urban canopy level was reduced
when roofs are applied, generating stronger (weaker) wind speed in higher
(lower) levels. The study by Owinoh et al. (2005) and Zilitinkevich et al.
(2006) shows a detailed discussion of the dynamics. For relative humidity, it
is increased due to the reduced temperature which is capable of lower the sat-
uration vapor pressure in CR and SPR. The increase in relative humidity in GR
canbeattributedtohigherevaporationandlowerhorizontalwindspeed.
These results indicate that advanced roofing technology mainly miti-
gates the air temperature but may exacerbate relative humidity. This exac-
erbation is related to effects induced by changes in vertical mixing. Studies
have shown that after the deployment of cool roofsand green roofs, the sen-
sible heat flux decreases, which in turn increases the stability of the atmo-
sphere and hence decreases the vertical mixing over the city (Li et al.,
2014;Sharma et al., 2016). When vertical mixing weakens, it takes more
time to establish the internal boundary layer when air flows develop from
rural to urban regions.Hence, the atmosphere is less affected by the surface
but more influenced by advection coming from upwind surfaces. Conse-
quently, the weaker mixing has a stronger advective effect on urban re-
gions. Finally, the stronger moisture advection in the air from rural
regions induces an increase in humidity over urban areas after installation
of advanced roofs (Li et al., 2014;Sharma et al., 2016).
3.5. Surface energy balance
Fig. 10 displays 6-day averaged diurnal cycle of surface latent heat flux
(Fig. 10a), sensible heat flux (Fig. 10b), net shortwave radiation (Fig. 10c),
net longwave radiation (Fig. 10d), and ground heat flux (Fig. 10e) from all
experiments. Positive values indicate incoming heat flux to the surface,
while negative values indicate outgoing heat flux to the atmosphere.
Fig. 10a shows that the CR and SPR decrease the latent heat flux, with
the peak reduced by 50 W m
−2
and 30 W m
−2
, respectively. The GR, on
the other hand, decreases the sensible heat flux by building the shade and
re-separates the available energy to increase latent heat flux by 80 W m
−2
via evapotranspiration, hence, reducing the temperature. For sensible heat
flux (Fig. 10b), the CR reduces the peak of sensible heat flux the most
Fig. 7. Hours of heat index difference that are >32.2 °C (90 °F) during the heatwave event (August 24, 2021) for (a) CR, (b) GR, and (c) SPR over CMA.
Fig. 8. The 6-day averaged 2-m relative humidity differences for (a) CR, (b) GR, and (c) SPR. Values are percentages.
H. Tan et al. Science of the Total Environment xxx (xxxx) xxx
9
(155 W m
−2
), followed by GR (103 W m
−2
)andSPR(84Wm
−2
). Conse-
quently,theseroofsareabletolimitthetransmissionofheatintourbaninte-
rior and the broader environment by decreasing sensible heat flux, so that the
near-surface temperature and the cooling consumption are reduced. For net
radiative flux, the CR and SPR share a similar tendency on both shortwave
and longwave radiation, indicating that the reflective effect in CR and the ab-
sorption by photovoltaic cells in SPR are similar. The GR can receive slightly
more net shortwave radiation during afternoon hours. The ground heat flux
decreased in CR, GR, and SPR during daytime compared with CTL, indicating
less heat flows into the building.
3.6. Impacts on cooling energy demand
This section addresses the summertime regional impacts of cool roofs,
green roofs, and solar panel roofs deployment on AC energy consumption.
Fig. 11a–c shows the AC energy consumption differences (MW km
−2
)
between CTL and the three roof experiments (CR, GR, and SPR), and
Fig. 11d shows the diurnal cycle of CMA AC consumption (MW km
−2
).
Overall, all three roofing strategies reduce the AC energy consumption of
the entire CMA. In particular, CR significantly decreases the AC energy
consumption (Fig. 11a) because CR has the least heat transmission into
the urban canopy layer compared to GR and SPR, especially during the
daytime when the air temperature and the AC demand are both at the
highest. The daily average reduction of AC energy consumption is 16.6 %
by CR, 14.0 % by GR, and 7.6 % by SPR, similar to what Salamanca et al.
(2016) found over Phoenix and Tucson (13–14 % by CR and 8.7–11 % by
SPR). In contrast, Salamanca et al. (2016), who observed that partially cov-
ered solar panel roofs reduce AC consumption during the day but increase
cooling energy demand at night (because solar panels receive less incoming
solar radiation during the day, hence it permits less radiative cooling at
night), our study finds the AC consumption is reduced during both daytime
and nighttime over CMA by all roofing strategies (Fig. 11d). This may be
because more heat is conducted into the building from the rooftop during
the day and released at night when the rooftop is partially covered by
solar panels in Salamanca's study. Other possibilities also include various
city morphologies (i.e. building fraction, building sizes). Since Phoenix's
local cooling energy demand is higher than Chicago's (Hong et al., 2009),
the disparity in local cooling energy demand between CMA and Phoenix
may also be a significant contributing factor.
While the SPR shows the smallest savings on AC consumption, it gener-
ates electricity. A detailed description of electricity production by solar
photovoltaic panels is presented by Zonato et al. (2021). We assume the
conversion efficiency of the solar photovoltaic panels to be 0.19. The elec-
tricity production by solar panels starts at 0600 LST and peaks at 1300 LST
at 10.7 MW km
−2
, which almost overcomes the cooling energy demand
(orange line, Fig. 11d). If all the power generated by solar panel roofs is
used for cooling, we may expect a cooling energy savings of 46.7 % on a
daily average in the SPR experiment, which is considerably more than the
cooling energy savings in CR and GR.
4. Discussion and conclusions
This study evaluates the impacts of city-scale deployment of cool roofs,
green roofs, and rooftop solar panels on near-surface temperatures and
cooling energy demand using a fully coupled modeling system consisting
of an urban-resolving regional climate model during a summertime
heatwave event over CMA.The results of the study can informdevelopment
and implementation of sustainable approaches that decrease the direct
effects of urbanization, lower summertime cooling energy demand, and
help minimize greenhouse gas emissions in the long term over CMA.
We find that the distribution of cool roofs, green roofs, and solar panel
roofs reduces the near-surface temperature and AC consumption demand
at the city scale, especially during daytime when the air temperature is
the highest. In particular, cool roofs can reduce the near-surface tempera-
ture by 1.5 °C, followed by 1.2 °C for green roofs and 0.6 °C for solar
panel roofs. In line with Georgescu et al. (2014) and Sharma et al. (2016),
green roofs have a slight nighttime warming effect because the rooftop
soil layers accumulate extra solar radiation during the daytime and
discharge it at night. The peak reduction in the daily urban heat island is
3.9 °C for cool roofs, 2.5 °C for green roofs, and 2.3 °C for solar panel
roofs. Examination of the effect of urban fraction was revealing: The cool
roof experiment was the only one where the 2-m temperature trended
Fig. 9. Vertical profile differences between CTL and CR, GR, and SPR on (a–c) temperature (°C), (d–f) horizontal wind (m s
−1
), and (g–i) relative humidity (%) for CMA-
averaged from August 21–27th. Black box indicates the heatwave event (Aug 24th).
H. Tan et al. Science of the Total Environment xxx (xxxx) xxx
10
down as the urban fraction increased, indicating the cool roof is the most
effective mitigation strategy among these three rooftop options. The reduc-
tion in heat index is relatively small, especially for green roofs because of
the increase in humidity. The peak daily heat island mitigations based on
2-m temperature averaged is over LCZ 8, which does not own the largest
urban fraction. However, the idealized two-dimensional experiments by
Zonato et al. (2021) show that the mitigation effect from roofing on air tem-
perature varies almost linearly with the building surface to total surface
fraction during summer time. The mitigation effect is also higher for low
buildings, with a nonlinear decrease in the impact with building heights.
Additionally, they draw the conclusion that the urban configuration with
the lowest buildings and the highest building area to total area ratio
shows the highest effect of the roofing mitigation. We will keep investigat-
ing this by designing more experiments based on different urban fractions
and building heights in real-time simulation in the future. The deployment
of the roof surface also influences the structure of the boundary layer during
the heatwave event. The temperature, horizontal winds, and relative
humidity in the lower atmosphere alter due to the mitigation of UHI from
rooftops.
Because all the roofing strategies offer cooling effects, they reduce AC
consumption. For example, the citywide cooling energy consumption can
be reduced by 16.6 %, 14.0 %, and 7.6 %, when cool roofs, green roofs,
and solar panel roofs are deployed, respectively. Although solar panel
roofs show the smallest reduction in energy consumption, if we assume
that all electricity production can be applied to cooling demand, we can
expect almost a savings of almost half (46.7 %) on cooling energy demand.
Several trade-offs must be considered when choosing between cool
roofs, green roofs, and solar panel roofs as artificial mitigation techniques.
The large-scale deployment of cool roofs has the best potential for cooling
effects and cooling energy saving; they cost less than the other two technol-
ogies; and they do not require additional water. However, the performance
of cool roofs can be misleading in perturbation experiments because the
Fig. 10. Diurnal cycle of energy balance terms for August 21–27th, 2021: (a) latent heat flux, (b) sensible heat flux, (c) net shortwave radiation, (d) net longwave radiation,
and (e) ground heat flux for CTL (black), Cool Roof (blue), Green Roof (green), and Solar Panel Roof (red).
H. Tan et al. Science of the Total Environment xxx (xxxx) xxx
11
roof albedo can decrease over time because of weathering and the accumu-
lation of dirt on the surface (Bretz and Akbari, 1997). Green roofs increase
water use and local humidity, but they can add a considerable amount
of vegetation to the urban environment, which has several potential co-
benefits, and they can reduce pollutants through dry deposition and absorp-
tion of gaseous pollutants through leaf stomata (Yang et al., 2008;Nowak
et al., 2014). However, purely from a heat mitigation perspective, we
should also consider the warming effects such roofs have at night over
urban areas (Georgescu et al., 2014;Sharma et al., 2016) and their minimal
impact on reducing theheat index. Regarding the heat index, it is important
to note that the current version of our model does not explicitly represent
street vegetation, which can further mitigate heat exposure through shad-
ing at local scales (Mussetti et al., 2020). Solar photovoltaic panels are gen-
erally discussed in the context of reduced reliance on fossil fuels and hence
global warming instead of in the context of local-scale heat mitigation.
However, this study demonstrates that solar panels can be another effective
mitigation method in the long term, thanks to their ability to generate
considerable amountsof electricity, whichcan indirectly reduce net cooling
energy demands. The ongoing scientific development in improving solar
conversion efficiency of the panels may further add to this benefit in the
future. It is important to also consider the trade-offs between urban heat
mitigation and air quality (Li et al., 2014;Sharma et al., 2016), which we
do not address in the present study.
Although we show that advanced roofing technologies are a productive
way to mitigate urban temperature during a summer heatwave event, the
efficacy of these mitigation methods would depend on the time of the
year and is a function of the simplifications in the urban model. All urban
characteristics (morphological, thermodynamic, radiative, aerodynamic,
etc.) are difficult to accurately prescribe even at higher resolutions because
of uncertainties in both observations and model parameterizations (Chen
et al., 2012;Chakraborty et al., 2021;Qian et al., 2022). For instance, a
city may possess a distinct morphology (i.e., building fraction, building
sizes, etc.), whose impacts on urban micro-climate can deviate from those
simulated by bulk parameterizations in the models, thus modulating the
quantitative benefits derived for these rooftop strategies. A broader clima-
tological study accounting for various seasons would require enormous
computational time at 500-m resolution; such a study is beyond the scope
of the current study but would be a consideration in the future. Another
caveat from the model perspective is that although in BEP + BEM the
anthropogenic heat flux is triggered by air conditioning, the heat exchanges
between interior building and outside air, and heat release by equipment
and people within the building, it doesn't include anthropogenic heat
from traffic and industrial activities. We plan to conduct more tests in
BEP + BEM with more realistic anthropogenic heat fluxes as the next step.
CRediT authorship contribution statement
Haochen Tan: Visualization; Investigation; Writing, Reviewing and
Editing; Validation; Formal analysis.
Rao Kotamarthi: Conceptualization, Methodology, Funding acquisition.
Jiali Wang: Methodology; Reviewing and Editing.
Yun Qian: Supervision; Project administration;
TC Chakraborty: Investigation; Writing; Validation.
Data availability
Data will be made available on request.
Declaration of competing interest
The authors declare that they have no known competing financial inter-
ests or personal relationships that could have appeared to influence the
work reported in this paper.
Acknowledgments
This study is supported by COMPASS-GLM, a multi-institutional project
supported by the U.S. Department of Energy (DOE), Office of Science,
Fig. 11. Daytime air-conditioning electricity consumption change (compared to CTL) due to (a) Cool Roof, (b) Green Roof, and (c) Solar Panel Roof over 6-day extreme heat
period (August 21–27, 2021) across CMA. (d) Diurnal cycle of simulated air-conditioning electricity consumption for CTL (black), cool roof (blue), green roof (green),and
solar panel roof (red) and the electricity production generated by solar panel roof (EP by SPR, orange). Unit in MW km
−2
.
H. Tan et al. Science of the Total Environment xxx (xxxx) xxx
12
Office of Biological and Environmental Research as part of the Regional and
Global Modeling and Analysis (RGMA) program, Multi-sector Dynamics
Modeling (MSD) program, and Earth System Model Development (ESMD)
program. We also acknowledge support from the CROCUS project funded
by DOE BER under contract number DE-FOA-0002581. Computational re-
sources are provided by the DOE-supported National Energy Research Sci-
entific Computing Center and Argonne Leadership Computing Facility. All
the calculations are done using the NCAR Command Language (version
6.6.2) (2019, Boulder, Colorado: UCAR/NCAR/CISL/TDD, https://doi.
org/10.5065/D6WD3XH5). MesoWest datasets were used in this study
(http://mesowest.utah.edu).
References
Akbari, H., Menon, S., Rosenfeld, A., 2009. Global cooling: increasing world-wide urban albedos
to offset CO2. Clim. Chang. 94, 275–286. https://doi.org/10.1007/s10584-008-9515-9.
Bou-Zeid, E., Meneveau, C., Parlange, M.B., 2004. Large-eddy simulation of neutral atmo-
spheric boundary layer flow over heterogeneous surfaces: blending height and effective
surface roughness. Water Reso ur. Res. 40, W025 05. https://do i.org/10.1029/
2003WR002475.
Bou-Zeid, E., Parlange, M.B., Meneveau, C., 2007. On the parameterization of surface
roughness at regional scales. J. Atmos. Sci. 64, 216–227. https://doi.org/10.1175/
JAS3826.1.
Bretz, S.E., Akbari, H., 1997. Long-term performance of high-albedo roof coatings. Energy.
Build. 25 (2), 159–167. https://doi.org/10.1016/S0378-7788(96)01005-5.
Brousse, O.,Martilli, A., Foley,M., Mills, G., Bechtel,B., 2016. WUDAPT, an efficient land use
producingdata tool for mesoscale models? Integrationof urban LCZ in WRF over Madrid.
Urban Clim. 17, 116–134. https://doi.org/10.1016/j.uclim.2016.04.001.
Burian, S.J., Brown, M.J., Linger, S.P., 2002. Morphological Analyses Using 3DBuilding Data-
bases. 74. Los Alamos National Laboratory, Los Angeles, California.
Burian, S.J., Brown, M.J., Linger, S.P., 2002. Morphological Analyses Using 3DBuilding Data-
bases, p. 74 Los Angeles, California.
Chakraborty, T.C., Lee, X., Ermida, S., Zhan, W., 2021. On the land emissivity assumption and
Landsat-derived surface urban heat islands: a global analysis. Remote Sens. Environ. 265,
112682. https://doi.org/10.1016/j.rse.2021.112682.
Chen, F., Dudhia, J., 2001. Coupling an advanced land surface–hydrology model with the
Penn State–NCAR MM5 modeling system. Part I: model implementation and sensitivity.
Mon. Weather Rev. 129 (4), 569–585.
Chen, F., Bornstein, R., Grimmond, S., Li, J., Liang, X., Martilli, A., Miao, S., Voogt, J., Wang,
Y., 2012. Research priorities in observing and modeling urban weather and climate. Bull.
Am. Meteorol. Soc. 93 (11), 1725–1728. https://doi.org/10.1175/BAMS-D-11-00217.1.
Chen, F., Yang, X., Zhu, W., 2014. WRF simulations of urban heat island under hot-weather
synoptic condi tions: the case study of Hangzhou Ci ty, China. Atmo s. Res. 138 (1),
364–377 doi:j.atmosres.2013.12.005.
Chen, K., Newman, A.J., Huang, M., Coon, C., Darrow, L.A., Strickland, M.J., Holmes, H.A.,
2022. Estimating heat-related exposures and urban heat island impacts: a case study for
the 2012 Chicago heatwave. GeoHealth 6, e2021GH000535. https://doi.org/10.1029/
2021GH000535.
Chen, L., Ng, E., 2011. Quantitative urban climate mapping based on a geographical database:
a simulation approach using Hong Kong as a case study. Int. J. Appl. Earth Obs. Geoinf.
13 (4), 586e594.
Ching, J., Mills, G., Bechtel, B., See, L., Feddema, J., Wang, X., Ren, C., Brousse, O., Martilli,
A., Neophytou, M. , Mouzourides, P. , Stewart, I., Hann a, A., Ng, E., Foley, M. ,
Alexander, P., Aliaga, D., Niyogi, D., Shreevastava, A., Bhalachandran, P., Masson, V.,
Hidalgo, J., Fung, J., Andrade, M., Baklanov, A., Dai, W., Milcinski, G., Demuzere, M.,
Brunsell, N., Pesaresi, M., Miao, S., Mu, Q., Chen, F., Theeuwes, N., 2018. WUDAPT: an
urban weather, cl imate, and environ mental modeling inf rastructure fo r the
anthropocene. Bull. Am. Meteorol. Soc. 99 (9), 1907–1924. https://doi.org/10.1175/
BAMS-D-16-0236.1.
Davies, F., Middleton, D.R., Bozier, K.E., 2007. Urban air pollution modelling and measure-
ments of boundary layer height. Atmos. Environ. 41, 4040–4049. https://doi.org/10.
1016/j.atmosenv.2007.01.015.
de Munck, C.S., Lem onsu, A., Bouzouidja, R., Masson , V., Claverie, R., 20 13. The
GREENROOF module (v7.3) for modelling green roof hydrological and energetic perfor-
mances within TEB. Geosci. Model Dev. 6 (6), 1941–1960. https://doi.org/10.5194/
gmd-6-19 41-2013.
Demuzere, M., Hankey, S., Mills, G., Zhang, W.W., Lu, T.J., Bechtel, B., 2020. Combining ex-
pert and crowd-sourcedtraining data to mapurban form and functions for the continental
US. Sci. Data 7, 264. https://doi.org/10.6084/m9.figshare.12599966.
Demuzere, M., Argüeso, D., Zonato, A., Kittner, J., 2022. W2W: a python package that injects
WUDAPT's local climate zone information in WRF. J. Open Source Softw. 7 (76), 4432.
https://doi.org/10.21105/joss.04432.
Dominguez, A., Kleissl, J., Luvall, J.C., 2011. Effects of solar photovol taic panels on roof
heat transfer. Sol. Energy 85, 2244–2255. https://doi.org/10.1016/j.solener.2011.
06.010.
Dudhia, J., 1989. Numerical study of convection observed during the winter monsoon exper-
iment using a mesoscale two-dimensional model. J. Atmos. Sci. 46, 3077–3107.
Ek, M.B., Mitchell, K.E., Lin, Y., Rogers, E., Grunmann, P., Koren, V., Gayno, G., Tarpley, J.D.,
2003. Implementation of Noah land surface model advances in the National Centers for
Environmental Prediction operat ional mesoscale Eta model. J. Geophys. Res. Atmos.
108 (2003 ). https://doi.org/10.1029/2002jd003296.
Foley, M., 2015. Chicago WUDAPT Level 0 training data is licensed under CC-BY-NC-SA 4.0.
Available at www.wudapt.org/cities/raw-training-area-download/.
Georgescu, M., 20 15. Challenges associated with adaptation to future urban expans ion.
J. Clim. 28, 2544–2563. https://doi.org/10.1175/JCLI-D-14-00290.1.
Georgescu, M., Moustaoui, M., Mahalov, A., Dudhia, J., 2013. Summer-time climate impacts
of projected megapolitan expansion in Arizona. Nat. Clim. Change 3, 37–41. https://doi.
org/10.1038/nclimate1656.
Georgescu, M., Morefield, P.E., Bierwagen, B.G., Weaver, C.P., 2014. Urban adaptation can
roll back warming of emerging megapolitan regions. Proc. Natl. Acad. Sci. U. S. A. 111,
2909–2914. https://doi.org/10.1073/pnas.1322280111.
Giovannini, L. , Zardi, D., de Fra nceschi, M., Chen, F., 2014. Numer ical simulatio ns of
boundary-lay er processes and ur ban-induced al terations in an alpine valley. Int.
J. Climatol. 34 (4), 1111–1131. https://doi.org/10.1002/joc.3750.
Gutiérrez, E., Martilli, A., Santiago, J.L., González, J.E., 2015. A mechanical drag coefficient
formulation and urban canopyparameter assimilation technique for complex urbanenvi-
ronments. Bound.-Layer Meteorol. 157 (2), 333–341. https://doi.org/10.1007/s10546-
015-0051-7.
Hammerberg, K., Brousse, O., Martilli, A., Mahdavi, A., 2018. Implications of employing de-
tailed urban canopy parameters for mesoscale climate modelling: a comparison between
WUDAPT and GIS databases over Vienna, Austria. Int. J. Climatol. 38, e1241–e1257.
https://doi.org/10.1002/joc.5447.
Harris, L., Kotamarthi, V.R., 2005. The characteristics of the Chicago lake breeze and its ef-
fects on trace particle transport: results from an episodic event simulation. J. Appl. Me-
teor. 44, 1637–1654.
He, X., Li, Y., Wang, X., Chen, L., Yu, B., Zhang, Y., Miao, S., 2019. High-resolution dataset of
urban canopy parameters for Beijing and its application to the integrated WRF/urban
modelling system. J. Clean. Prod. 208, 373e383.
Hong, S.Y., Lim, J.O.J., 2006.The WRF single-moment 6-class microphysics scheme (WSM6).
J. Korean Meteor. Soc. 42, 129–151.
Hong, T.Z., Sartor, D.,Mathew, P., Yazdanian, M., 2009. Comparisonof HVAC simulations be-
tween EnergyPl us and DOE-2.2 for d ata centers. ASHRA E Transcations . https://
escholarship.org/content/qt2dq2w3b3/qt2dq2w3b3.pdf?t=lnqp8p.
Jaffal, I., Ouldboukhitine, S.E., Belarbi, R., 2012. A comprehensive study of the impact of
green roofs on building energy performance. Renew. Energy 43, 157–164. https://doi.
org/10.1016/j.renene.2011.12.004.
Janicke,B., Meier, F., Fenner,D., Fehrenbach, U.,Holtmann, A., Scherer, D., 2017.Urbanrural
differences in near-surface air temperature as resolved by the Central Europe refined
analysis (CER) : sensitivity to pl anetary boundary layer schemes and urban cano py
models. Int. J. Climatol. 37 (4), 2063–2079. https://doi.org/10.1002/joc.4835.
Janjic, Z.I., 2002. Nonsingular Implementation of the Mellor-YamadaLevel 2.5 Scheme in the
NCEP Mesomodel. NCEP Office Note. No. 437 61pp.
Kim, Y., Sartelet, K., Raut, J.C., Chazette, P., 2013. Evaluation of the weather research and
forecast/urban model over greater Paris. Bound.-Layer Met eorol. 149 (1), 105–132.
https://doi.org/10.1007/s10546-013-9838-6.
Kong, F., Yan, W., Zheng, G., Yin, H., Cavan, G., Zhang, W., Zhang, N., Cheng, L., 2016. Re-
trieval of three -dimensional tree canopy and s hade using terrestrial laser scanning
(TLS) data to analyze the cooling effect of vegetation. Agric. For. Meteor. 217, 22–34.
https://doi.org/10.1016/j.agrformet.2015.11.005.
Kumar, A., Chen, F., Barlage, M., Ek, M.B., Niyogi, D., 2014. Assessing impact of integrating
MODIS vegetation data in the weather research and forecasting (WRF) model coupled
to two different canopy-resistance approaches. J. Appl. Meteor. Clim. 53, 1362–1380.
https://doi.org/10.1175/JAMC-D-13-0247.1.
Kusaka, H., Kondo, H., Kikegawa, Y., Kimura, F., 2001. A simple single-layer urban canopy
model for atmospheric models: comparison with multi-layer and slab models. Bound.-
Layer Meteorol. 101, 329–358. https://doi.org/10.1023/A:1019207923078.
Lee, S.H., Kim, S.W ., Angevine, W.M. , Bianco, L., McKeen , S.A., Senff, C.J., Trainer, M.,
Tucker, S.C., Zamora, R.J., 2011. Evaluation of urban surface parameterizations in the
WRF model using measurements during the Texas air quality study 2006 field campaign.
Atmos. Chem. Phys. 11 (5), 2127–2143. https://doi.org/10.5194/acp-11-2127-2011.
Li, D., Bou-Zeid, E., Oppenheimer, M., 2014. The effectiveness of cool and green roofs as
urban heat island mitigation strategies. Environ. Res. Lett. 9, 055002. https://doi.org/
10.1088/ 1748-932 6/9/5/0550 02.
Li, H., Zhou, Y., Wang, X., Zhou, X., Zhang, H., Sodoudi, S., 2019. Quantifying urban heat is-
land intensity and its physical mechanism using WRF/UCM. Sci. Total. Environ. 650,
3110–3119. https://doi.org/10.1016/j.scitotenv.2018.10.025.
Li, X.-X., Norford, L.K., 2016. Evaluation of cool roof and vegetations in mitigating urban heat
island in a tropical city, Singapore. Urban. Clim. 16, 59–74. https://doi.org/10.1016/j.
uclim.2015.12.002.
Liao, W., Liu, X., Li, D., Luo, M., Wang, D., Wang, S., Baldwin, J., Lin, L., Li, X., Feng, K.,
Hubacek, H., Yang, X., 2018. Stronger contributions of urbanization to heat wave trends
in wet climates. Geophys. Res. Lett. 45 (20), 11310–11317. https://doi.org/10.1029/
2018gl079679.
Livezey, R.E., Tinker, R., 1996. Some meteorological, climatological, and microclimatological
considerations of the severe US heat wave of mid-July 1995. Bull. Am. Meteorol. Soc. 77,
2043–2054. https://doi.org/10.1175/1520-0477(1996)077<2043:SMCAMC>2.0.CO;2.
Ma, S., Goldstein, M., Pitman, A., Haghdadi, N., MacGill, I., 2017. Pricing the urban cooling
benefits of solar panel deployment in Sydney, Australia. Sci Rep. 7, 43938. https://doi.
org/10.1038/srep43938.
Martilli, A., Clappier, A., Rotach, M.W., 2002. An urban surface exchange parameterisation
for mesoscale models. Bound.-Layer Meteorol. 104, 261–304. https://doi.org/10.1023/
A:1016099921195.
Masson, V., Bonhomme, M., Salagnac, J.-L., Briottet, X., Lemonsu, A., 2014. Solar panels re-
duce both global warming and urban heat island. Front. Environ. Sci. 2, 1–10. https://
doi.org/10.3389/fenvs.2014.00014.
Mavrogianni, A., Davies, M., Batty, M., Belcher, S.E., Bohnenste ngel, S.I., Carruthers, D.,
Chalabi, Z., Croxford, B., Demanuele, C., Evans, S., Giridharan, R., Ye, Z., 2011. The
H. Tan et al. Science of the Total Environment xxx (xxxx) xxx
13
comfort, energy and health implications of London’s urban heat island. Build. Serv. Eng.
Res. Technol. 32, 35–52. https://doi.org/10.1177/0143624410394530.
McRae, I.,Freedman, F., Rivera, A., Li, X.W., Dou, J.J.,Cruz, I., Ren, C., Dronova, I., Fraker, H.,
Bornstein, R., 2020. Integration of the WUDAPT, WRF, and ENVI-met models to simulate
extreme daytime temperature mitigation strategies in San Jose, California. Build. Envi-
ron. 184, 107180. https://doi.org/10.1016/j.buildenv.2020.107180.
Millstein, D., Menon, S., 2011. Regional climate consequences of large-scale cool roof and
photovoltaic ar ray deployment. Environ. Res. Lett. 6, 034001. http s://doi.org/1 0.
1088/1748-9326/6/3/034001.
Mlawer, Eli, J., Taubman, S.J., Brown, P.D., Iacono, M.J., Clough, S.A., 1997. Radiative trans-
fer for inhomogeneous atmospher es: RRTM, a validated correlated- k model for the
longwave. J. Geophys. Res. 102, 16663–16682. https://doi.org/10.1029/97JD00237.
Mohan, M., Gupta, A., Bhati, S., 201 4. A modifed appro ach to analyze the rmal comfort
classifcation. Atmos. Clim. Sci. 4, 13. https://doi.org/10.4236/acs.2014.41002.
Monin, A.S., Obukhov, A.M., 1954. Basic Laws of turbulent mixing in the surface layer of the
atmosphere. Contrib. Geophys. Inst. Acad. Sci. USSR 24, 163–187.
Morini, E., Touchaei, A.G., Rossi, F., Cotana, F., Akbari, H., 2017. Evaluation of albedo en-
hancement to mitigate impacts of urban heat island in Rome (Italy) using WRF meteoro-
logical model. Urban Clim. 24, 551–566. https://doi.org/10.1016/j.uclim.2017.08.001.
Mussetti, G., Brunner, D., Henne, S., Allegrini, J., Krayenhoff, E.S., Schubert, S., Feigenwinter,
C., Vogt, R., Wicki, A., Carmeliet, J., 2020. COSMO-BEP-tree v1.0: a coupled urban cli-
mate model with ex plicit represen tation of street trees. G eosci. Model Dev . 13,
1685–1710. https://doi.org/10.5194/gmd-13-1685-2020.
Nowak, D.J., Hirabayashi, S., Bodine, A., Greenfield, E., 2014. Tree and forest effects on air
quality and human health in the United States. Environ. Pollut. 193, 119–129. https://
doi.org/10.1016/j.envpol.2014.05.028.
Owinoh, A.Z., Hunt, J.C., Orr, A., Clark, P., Klein, R.,Fernando, H.J.S.,Nieuwstadt, F.T.,2005.
Effect of changing surface heatflux on atmospheric boundary-layer flowover flat terrain.
Bound.-Layer Meteor. 116, 331–361.
Patel, P., Karmakar, S., Ghosh, S., Niyogi, Dev, 2020. Improved simulation of veryheavy rain-
fall eventsby incorporating WUDAPTurban land use/landcover in WRF. Urban Clim. 32,
100616. https://doi.org/10.1016/j.uclim.2020.100616.
Poulos, G.S., Blumen, W., Fritts, D.C., Lundquist, J.K., Sun, J.L., Burns, S.P., Nappo, C., Banta, R.,
Newsom, R., Cuxart, J., Terradellas, E., Balsley, B., Jensen, M., 2002. CASES-99: a compre-
hensive investigation of the stable nocturnal boundary layer. Bull. Amer. Meteor. Soc. 83,
555–581. https://doi.org/10.1175/1520-0477(2002)083,0555:CACIOT.2.3.CO;2.
Qian, Y., Chakraborty, T.C., Li, J., Li, D., He, C., Sarangi, C., Chen, F., Yang, X., Leung, L.R.,
2022. Urbanization impacton regional climate and extreme weather:current understand-
ing, uncertainties, and future research directions. Adv. Atmos. Sci. 39, 819–860. https://
doi.org/10.1007/s00376-021-1371-9.
Ribeiro,I., Martilli, A., Falls, M., Zonato, A., Villaba,G., 2020. Highly resolved WRF-BEP/BEM
simulations over Barcelona urban area with LCZ. Atmos. Res. 248, 105220. https://doi.
org/10.1016/j.atmosres.2020.105220.
Ricard, S., Badia, A., Ventura, S., Gilabert, J., Martilli, A., Villalba, G., 2021. Sensitivity study
of PBL schemes and soil initialization using the WRF-BEP-BEM model over a Mediterra-
nean coastal cit y. Urban Clim. 39, 10 0982. https:// doi.org/10.10 16/j.uclim.2021.
100982.
Rothfusz, L.P., 1990. The heat index equation (or, more than you ever wanted to know about
heat index). Tech. Attachment, SR/SSD 90-23, U.S. National Weather Service, Southern
Regional Headq uarters, Forth Wo rth, TX. https://www. weather.gov/m edia/fc/ta_
htindx.PDF.
Salamanca, F., Martilli, A., 2010. A new building energy model coupled with an urban canopy
parameterization for urban climate simulations—part II. Validation with one dimention
off-line simulations. Theor. Appl. Climatol. 99, 345. https://doi.org/10.1007/s00704-
009-0143-8.
Salamanca, F., Krpo, A.,Martilli, A., Clappier, A., 2010. A new building energy model coupled
with an urban canopy parameterization for urban climate simulations - part I. Formula-
tion, verification, and sensitivity analysis of the model. Theor. Appl. Climatol. 99 (3–4),
331–344. https://doi.org/10.1007/s00704-009-0142-9.
Salamanca, F., Martilli, A., Tewari, M., Chen, F., 2012. A study of the urban boundary layer
using different urban parameterizations and high-resolution urban canopy parameters
with WRF. J. Appl. Meteorol. Climatol . 50, 1107–1128. https://doi.org/10.1175/
2010JAMC2538.1.
Salamanca, F., Tonse, S., Menon, S., Garg, V., Singh, K.P., Naja, M., Fischer, M.L., 2012a. Top-
of-atmosphere radiative cooling with white roofs: experimental verification and model-
based evaluation. Environ. Res. Lett. 7, 044007. https://doi.org/10.1088/1748-9326/
7/4/0440 07.
Salamanca, F., Georgescu, M., Mahalov, A., Moustaoui, M., Martilli, A., 2016. Citywide im-
pacts of cool roof and rooftop solar photovoltaic deployment on near-surface air temper-
ature and cooling energy demand. Bound.-Layer Meteorol. 161, 203–221. https://doi.
org/10.1007/s10546-016-0160-y.
Santamouris, M., 2014. Cooling the cities—Areviewofreflective and green roof mitigation
technologies to fight heatisland and improve comfort in urban environments. Sol. Energy
103, 682–703. https://doi.org/10.1016/j.solener.2012.07.003.
Santiago, J.L., Martilli, A., 2010. A dynamic urban canopy parameterization for mesoscale
models based on computational fluid dynamics reynoldsaveraged Navier–Stokes micro-
scale simulations. Bound.-Layer Meteorol. 137 (3), 417–439. https://doi.org/10.1007/
s10546-0 10-9538- 4.
Scherba, A., Sailor, D.J., Rosenstiel, T.N., Wamser, C.C., 2011. Modeling impacts of roof re-
flectivity, integrated photovoltaic panels and green roof systems on sensible heat flux
into the urban environment. Build. Environ. 46 (12), 2542–2551. https://doi.org/10.
1016/j.buildenv.2011.06.012.
Sharma, A., Conry, P., Fernando, H.J.S., Alan, F.H., Hellmann, J.J., Chen, F., 2016. Green and
cool roofs to mitigate urban heat island effects in the Chicago metropolitan area: evalua-
tion with a regional climate model. Environ. Res. Lett. 11 (6), 064004. https://doi.org/
10.1088/ 1748-932 6/11/6/064 004.
Skamarock, W.C., Klemp, J.B., Dudhia, J., Gill, D.O., Liu, Z., Berner, J., Wang, W., Powers, J.G.,
Duda, M.G., Barker, D.M., Huang, X.-Y., 2021. A Description of the Advanced Research WRF
Version4.3,(No.NCAR/TN-556+STR). https://doi.org/10.5065/1dfh-6p97.
Steadman, R.G., 1979. The assessment of sultriness. Part I: a temperature-humidity index
based on human ph ysiology and clo thing science. J . Appl. Meteor. 18, 861–873.
https://doi.org/10.1175/1520-0450(1979)018<0861:TAOSPI>2.0.CO;2.
Stewart, I.D., Oke, T.R., 2012. Local climate zones for urban temperature studies. Bull. Am.
Meteorol. Soc. 93 (12), 1879–1900. https://doi.org/10.1175/BAMS-D-11-00019.1.
Sun, T., Bou-Zeid, E., Wang, Z.-H., Zerba, E., Ni, G.-H., 2013. Hydrometeorological determi-
nants of green roof performance via a vertically-resolved model for heat and water trans-
port. Build. Environ. 60, 211–224. https://doi.org/10.1016/j.buildenv.2012.10.018.
Sun, T., Bou-Zeid, E., Ni, G.-H., 2014. To irrigate or not to irrigate: analysis of green roof per-
formance via a vertically-resolved hygrothermal model. Build. Environ. 73, 127–137.
https://doi.org/10.1016/j.buildenv.2013.12.004.
Sun, T., Grimmond, C.S.B., Ni, G.-H., 2016.How do green roofs mitigate urban thermal stress
under heat waves? J. Geo. Res: Atmos. 121, 5320–5335. https://doi.org/10.1002/
2016JD024873.
Synnefa, A., Dandou, A., Santamouris, M., Tombrou, M., Soulakellis, N., 2008. On the use of
cool materials as a heat islan d mitigation stra tegy. J. Appl. Meteo rol. Clim. 47,
2846–2856. https://doi.org/10.11 75/2008J AMC1830. 1.
Taha, H., 2013. Thepotential for air-temperature impact from large-scale deploymentof solar
photovoltaic arrays in urban areas. Sol. Energy 91, 358–367. https://doi.org/10.1016/j.
solener.2012.09.014.
Tan, H., Ray, P., Tewari, M., Brownlee, J., Ravindran, A., 2019. Response of near-surface me-
teorological conditions to ddvection under impact of the green roof. Atmosphere 10, 759.
https://doi.org/10.3390/atmos10120759.
Wang, X., Li, H., Sodoudi, S., 2022. The effectiveness of cool and green roofs in mitigating
urban heat island and improving human thermal comfort. Build. Environ. 217, 109082.
https://doi.org/10.1016/j.buildenv.2022.109082.
Wong, N.H., Chen, Y., 2005. Study of green areas and urban heat island in a tropical city.
Habit. Int. 29 (3), 547–558. https://doi.org/10.1016/j.habitatint.2004.04.008.
Yang, J., Yu, Q., Gong, P., 2008. Quantifying air pollution removal by green roofs in Chicago.
Atmos. Environ. 42, 7266–7273. https://doi.org/10.1016/j.atmosenv.2008.07.003.
Yang, J.C., Wang, Z.-H., Georgescu, M., Chen, F., Tewari, M., 2016. Assessing the impact of
enhanced hydrologicalprocesses on urban hydrometeorology with application to two cit-
ies in contrasting climates. J. Hydrometeorol. 17 (4), 1031–1047. https://doi.org/10.
1175/JHM-D-15-0112.1.
Yip, F.Y., Flanders, W.D., Wolkin, A., Engelthaler,D., Humble, W., Neri, A., Lewis, L., Backer,
L., Rubin, C., 2008. The impact of ex cess heat events in Maricopa Coun ty, Arizona:
2000–2005. Int. J. Biometeorol. 52, 765 –772. https://doi.org/10.1007/s 00484-008-
0169-0.
Zhang, J.C., Li, Y., Tao, W., Liu, J.F., Levinson, R., Mohegh, A., Ban-Weiss, G., 2019. Investi-
gating the urban air quality effects of cool wallsand cool roofs in southern California. En-
viron. Sci. Technol. 53 (13), 7532–7542. https://doi.org/10.1021/acs.est.9b00626.
Zhang, N., Chen, Y., Luo, Ling, Wang, Y.W., 2017. Effectiveness of different urban Heat Island
mitigation methods and their regional impacts. J. Hydro. 18 (11), 2991–3012. https://
doi.org/10.1175/JHM-D-17-0049.1.
Zilitinkevich, S.S., Hunt, J.C.R., Esau, I.N., Grachev, A.A., Lalas, D.P., Akylas, E., Joffre, S.M.,
2006. The influence of largeconvective eddies on the surface-layer turbulence. Q. J. R.
MeteorolSoc. 132, 1426–1456.
Zonato, A., Martilli, A., Sabatino, S.Di., Zardi, D., Giovannini, L., 2020. Evaluating the perfor-
mance of a novel WUDAPTaveraging technique to define urban morpholo gy with mes o-
scale models. Urban Clim. 31, 100584. https://doi.org/10.1016/j.uclim.2020.100584.
Zonato, A., Martilli, A., Gutierrez, E., Chen, F., He, C., Barlage, M., Zardi, D., Giovannini, L.,
2021. Exploring the effects of rooftop mitigation strategies on urban temperature and en-
ergy consumption. J. Geophys. Res. Atmos. 126, e2021JD035002. https://doi.org/10.
1029/2021JD035002.
H. Tan et al. Science of the Total Environment xxx (xxxx) xxx
14
