Building and Environment 217 (2022) 109061

Contents lists available at ScienceDirect

Building and Environment

journal homepage: www.elsevier.com/locate/buildenv

Experimental investigation of summer thermal performance of the green

roof system with mineral wool substrate
Danka Kostadinović a, *, Marina Jovanović a, Vukman Bakić a, Nenad Stepanić a, Maja Todorović b
a
Department of Thermal Engineering and Energy, “Vinča" Institute of Nuclear Sciences-National Institute of the Republic of Serbia, University of Belgrade, Mike Petrovića
Alasa 12-14, Po. Box. 522, 11000, Belgrade, Serbia
b
Faculty of Mechanical Engineering, University of Belgrade, Kraljice Marije 16, 11000, Belgrade, Serbia

A R T I C L E I N F O A B S T R A C T

Keywords: Green roof systems have become an important part in creating sustainable cities. They can provide a wide range
Cooling effect of economic, environmental, and social benefits. The goal of this research was to quantify the thermal perfor­
Green roof mance improvements from a green roof with mineral wool substrate installed on a school building in a humid
Heat flux
subtropical climate. In-situ measurements during a summer period included heat fluxes through the green and
Pearson correlation
Thermal performance
reference roof, vertical temperature profile through both roofs, the local air temperature above roofs, and local
Mineral wool meteorological parameters. Furthermore, the summer thermal performance of green and reference roof and the
green roof cooling effect were evaluated concerning meteorological parameters using the Pearson correlation
analysis. The results indicate that the green roof layers have improved thermal performance of the roof with
respect to reduced conductive heat flow by 57% and delayed heat transfer. The maximum and averaged refer­
ence roof to green roof outdoor surface temperature difference was 27.5 ◦ C and 5.5 ◦ C, respectively. It was found
that ambient temperature and relative humidity have a dominant role on the thermal performance of the green
and reference roof, while solar radiation and ambient temperature present the key meteorological determinants
of the green roof cooling effect.

cm) which can support a greater variety of plant species, even trees, and
1. Introduction can sustain walking paths and benches. However, they have higher
initial costs and require an irrigation system and more maintenance.
Over the last decades, the increased urbanization and economic The green roof substrate or growing medium is vitally important for
growth have resulted in less urban green spaces. Dense urban areas with successful vegetation growth on the green roof. It is usually a blend of
a large number of high-rise buildings can provide green space on roofs. soils, which needs to be lightweight, chemically stable, well-aerated,
Green roof systems have become an important part in creating sustain­ free-draining, and able to supply water and nutrients to plants [7].
able cities. They can provide a wide range of economic, environmental, The hydrophilic mineral wool in extensive green roof systems is used as
and social benefits which include the reduction of air pollution [1], a substitute for traditional soil substrate. Mineral wool is used as water
water run-off [2], urban heat island effects [3], noise [4], and an in­ retention and drainage layer as well as surface growth media. This type
crease in urban biodiversity [5]. Green roofs can significantly reduce of substrate rapidly absorbs rainwater and drains when material ap­
building cooling and heating energy demands through evapotranspira­ proaches saturation [8].
tion, photosynthesis, solar shading, and thermal insulation [6]. Many studies have been performed over the past years in different
There are two main classifications of green roofs: extensive green fields about green roofs. According to an experiment conducted in the
roofs and intensive green roofs, depending on weight, substrate layer, Mediterranean climate [9], a green roof reduces the roof surface tem­
plant type, maintenance, irrigation, and cost. Extensive green roofs have perature by an average of 12 ◦ C as compared to a black bituminous roof
a thin substrate layer (<15 cm) with a limited variety of plants. These in summer, while during the winter difference between bituminous and
green roof systems are lightweight, easy to install, and low maintenance. greens roofs was almost 4 ◦ C. In addition, the green roof reduces 100%
They are widely used in practice for retrofit public and commercial of thermal energy entering through the building’s roof in summer and
buildings. Intensive green roofs have a thicker substrate layer (20–200 reduces from 30% to 37% of the outgoing thermal energy in winter. The

* Corresponding author.
E-mail address: [email protected] (D. Kostadinović).

https://doi.org/10.1016/j.buildenv.2022.109061
Received 10 January 2022; Received in revised form 2 March 2022; Accepted 1 April 2022
Available online 5 April 2022
0360-1323/© 2022 Elsevier Ltd. All rights reserved.
D. Kostadinović et al. Building and Environment 217 (2022) 109061

roofs under free-floating conditions, while under air-conditioned con­

Nomenclature ditions correlation between solar radiation and heat flux through green
roof becomes insignificant. The most correlated factor of green roof’s
q heat flux [W/m2] temperatures was outdoor temperature, but also the least correlated one
Ts,in indoor roof surface temperature [◦ C] for green roof’s cooling effect. Another study from China [25] revealed
Ts,out outdoor roof surface temperature [◦ C] that solar radiation and relative humidity have the largest number of
Td drainage layer surface temperature [◦ C] significant and high correlation coefficients with thermal effect in­
Tgs growing substrate layer surface temperature [◦ C] dicators on the extensive and intensive green roof. The intensive green
Tv vegetation layer surface temperature [◦ C] roof was less sensitive to weather parameters in summer and winter
T10 local air temperature at 10 cm above roof [◦ C] compared to the extensive green roof.
Ta ambient temperature [◦ C] The previous studies clearly show differences in research findings
RH relative humidity [%] and demonstrated that thermal performance and effectiveness of green
SR solar radiation [W/m2] roof systems significantly depend on several parameters such as
WS wind speed [m/s] geographical location and meteorological conditions, substrate type and
thickness, water content, and vegetation type. To facilitate a more
Subscript massive use of green roofs more research is needed for a deeper un­
a ambient derstanding of their thermal behavior in different climates.
d drainage The lightweight green roofs could be good retrofitting solutions for
GR green roof existing building stock with limited load-bearing capacity. Despite this,
gs growing substrate experimental data have been lacking related to the thermal performance
in indoor and insulation effect of lightweight green roofs with the mineral wool
out outdoor substrate. Considering the widespread occurrence of old buildings, it is
RR reference roof important to conduct an in-depth experimental investigation of the
s surface thermal behavior of green roofs with the mineral wool substrate.
v vegetation While a few couple previous studies regard to green roofs with
mineral wool substrate were carried out during the winter season
implying brown and less developed vegetation, the summer season im­
plies green and fully developed vegetation cover. Based on previous
experimental study conducted in a Mediterranean climate [10] analysed
claims and conclusions this research focuses on the summer thermal
the different thermal impacts of extensive green roofs in summer and
performance of a green roof system with the latest generation rock
winter. It was shown that the thermal differences between green roofs
mineral wool substrate installed on a real-scale school building in
and standard roofing in the summer are much higher than in winter.
Serbia. This latest generation of rock mineral wool for green roofs is
Another study, in Athens, Greece showed that a green roof installed on a
formed by a mechanical interconnection of pure mineral fibres without
two-story commercial building reduced cooling and heating loads by
any additives or binders. At this point, Serbia has a small number of
18.7% and 11.4%, respectively [11]. In contrast, research in Hong Kong
green roofs and an absence of local experimental research related to
revealed that green roof had a heat-sink effect and increased downward
green roofs [26]. The lack of local performance data can be an obstacle
heat flux and indoor temperatures by 1–2 ◦ C [12].
for the government and investors in implementing green roofs. The
Some authors claim that the influence of the green roof system on the
objectives of this research are to:
building’s heating load was insignificant [13]. However, some studies
have found that the installation of green roofs increased building energy
(1) investigate the influence of green roof with the latest generation
consumption during winter [14,15]. Field measurements conducted
rock mineral wool growing substrate on the vertical temperature
during winter in Canada [16] showed that the green roof had lower heat
profile and heat flux through the roof,
loss compared to the conventional roof, while when the growing media
(2) analyse summer thermal behavior of green roof in comparison to
was frozen or covered with snow there was no difference in heat loss
a reference roof
between the green and conventional roof. Increasing substrate depth
(3) explore the influence of meteorological parameters on summer
results in reductions in heat flux into or out of the building [17,18].
thermal performance of green roof and reference roof and cooling
However, one study claim doubling the substrate depth from 7.5 cm to
effect of green roof.
15 cm had no impact in lowering net heat loss [16].
The results from the study in China [19] demonstrate that higher
The findings of a green roof experiment are applicable for the
water content contributed to improving the thermal performance of the
particular region where the experiment was performed. Knowledge and
green roof, while the study performed in Italy, found that higher water
expertise on the climate and environment where the green roof will be
content had a negative effect on the thermal performance of the green
located are crucial for building a highly efficient green roof. Some
roof [20]. The results of numerical analyses validated with in-situ ex­
quantitative data in the context of correlation between meteorological
periments [21] showed that the process of water freezing in wintertime
conditions and thermal performance and cooling effect of the light­
decreased the heat load of extensive green roofs with lightweight min­
weight green roofs are needed, as the results for green roofs with soil
eral wool substrate up to 30% due to latent heat storage. It was found,
substrate and from another climatic zone may not be directly applicable.
that the thermal effectiveness of an extensive green roof varies in
The results from this study will help understand the heat transfer
different climates and at different times of the day based on long-term
mechanisms associated with the lightweight green roof with the mineral
measurement [22]. The results of a study conducted on a tropical
wool substrate so the benefits of their use in humid subtropical climate
green roof have shown a one-day time lag of heat flux through the roof
could be determined.
after a long sunshine day. In addition, green roof block 43.9 TJ of solar
energy penetration into public housing buildings in one summer [23].
2. Methodology
The local meteorological conditions play a significant role in the
thermal and energy performances of green roofs. A field experiment and
2.1. Experimental site and climatic conditions
correlation analysis in Shanghai [24] showed that solar radiation is the
most correlated factor with heat flux through green and conventional
The experimental site is located in the New Belgrade municipality of

2
D. Kostadinović et al. Building and Environment 217 (2022) 109061

Belgrade (44◦ 49′ 14′′ N, 20◦ 27′ 44′′ E) the capital city of Serbia. New reference roof and green roof are 0.41 W/m2K and 0.27 W/m2K
Belgrade is one of the central urban business districts in Serbia and one respectively, which were calculated using the estimated method [28].
of the major ones in Southeast Europe as well as the second-most The layer of gravel was removed and a green roof was installed on PVC
populous municipality of Serbia. Belgrade lies at an average elevation foil. From bottom to top, the green roof consists of four layers including
of 117 m above sea level and has a humid subtropical type of climate. waterproof layer, drainage layer with a water reservoir, rock mineral
This climate is considered to be Cfa, according to the Köppen-Geiger wool as a growing substrate, and vegetation mat. The mineral wool
climate classification [27]. Belgrade has four seasons, with warm to hot growing substrate is a porous lightweight layer made of mainly volcanic
summers and cold winters. The warmest month is July with an average materials which represent good media for plants rooting and growing.
maximum temperature of 23 ◦ C. January is the coldest month, with an The vegetation mat consists of the following eight Sedum species: Sedum
average temperature of 1.4 ◦ C. Belgrade is moderately sunny with 2.112 acre, Album coral carpet sedum, Sedum wall album, Sedum hybridum,
average sunny hours during the year. Sedum sexangulare, Sedum hispanicum, Sedum floriferum, and Sedum
Kamtschaticum. The cross-sectional view of the green roof and reference
2.2. Experimental set-up roof experimental plot and position of sensors are shown in Fig. 2.
Below the green and reference roof, there are locker rooms that were
The field experiment was conducted in a school building built in not occupied due to school summer break. In both rooms, the indoor air
1975 with a three-story reinforced concrete structure. The selected temperature of 24 ◦ C was maintained by air conditioners. At the time of
building represents a typical school building with regards to the archi­ the measurement period, the green roofs had existed for more than a
tectural characteristics and roof type and material, from the period of year hence the roof was uniformly covered with vegetation and Sedum
most intensive construction of larger school buildings in urban areas of plants were well established. The experimental green roof has not an
Serbia. In November 2018, a lightweight extensive green roof with a irrigation system.
surface of 25 m2 was installed on the flat roof of a building wing at a
height of approximately 4 m from the ground level. The adjacent 2.3. Experimental methods
reference roof plot with the same surface area has been chosen to be
compared with the green one. Fig. 1 presents locations of the experi­ In this paper the empirical method was used, with quantitative data
mental green roof and the reference roof plots at the school building. collection through experimental measurements. A statistical analysis
From bottom to top, the structural roof consists of a concrete slab, was also conducted to explore the influence of meteorological parame­
vapor barrier, bitumoperlite, polyurethane, hydro isolation, and con­ ters on summer thermal performance of green roof and reference roof
crete slabs (30 × 30 cm) with a total thickness of 45 cm. The roof has a and green roof cooling effect. An experimental measurement was con­
5% slope for rainfall water management. In 2017, new hydro isolation ducted from 3 to 6 July 2020. Experimental equipment has been placed
was added on top of the roof which includes bitumen with fiberglass on both roof plots to compare the thermal performance of the green roof
mesh, PVC foil, and gravel with a total thickness of 5 cm. The U value for and the reference roof. Both roofs were equipped with the heat flux

Fig. 1. a) The experimental site located at the School building, Belgrade, Serbia. b) Locations of the green roof experimental plot (GR) and the reference roof plot
(RR) on the top of the School building. c) The experimental green roof plot with environmental monitoring sensors. d) The reference roof plot with environmental
monitoring sensors. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

3
D. Kostadinović et al. Building and Environment 217 (2022) 109061

high solar radiation and ambient temperature, clear sky condition, and
an absence of precipitations. The measured ambient temperature, rela­
tive humidity, and solar radiation are shown in Fig. 3, while wind speed
is shown in Fig. 4. During the hottest time of the day, the solar radiation
was around 900 W/m2, the ambient temperature was around 35 ◦ C, and
the relative humidity was between 35% and 45%. The wind speed
ranged from 0.2 m/s to 4.4 m/s.
The following parameters were compared on both roofs: outdoor and
indoor roof surface temperature, the local air temperature above roofs,
and heat flux through roofs. When it comes to the graphs of tempera­
tures, the given measured values are the means of measured values at
three measuring points. The first analysis compared the outdoor surface
temperature of the green roof and reference roof shown in Fig. 5. As
expected, the temperature of the outdoor surface of reference roof has
significantly higher values compared to the green roof and ambient
temperature, with large daily temperature fluctuations. It can be
observed that the highest reference roof-green roof surface temperature
differences are between 13:00 and 15:00. The maximum and averaged
Fig. 2. Schematic cross-sectional view of the green roof system and reference reference roof to green roof outdoor surface temperature difference was
roof with positions and heights of the installed sensor. (For interpretation of the
27.5 ◦ C and 5.5 ◦ C, respectively. Similar findings were reported in NYC
references to color in this figure legend, the reader is referred to the Web
[29], where daily maximum and averaged summer black roof to green
version of this article.)
roof membrane temperature difference was 33 ◦ C and 9.9 ◦ C, respec­
tively. This large temperature reduction is caused by solar shading,
sensor attached to the ceiling to record the heat flow through the green
thermal insulation, and latent heat loss through evapotranspiration. A
roof (qGR) and reference roof (qRR). To obtain temperature distribution
small amount of solar radiation is reflected from the reference roof,
along the vertical direction thermocouples have been installed at three
while a large part is absorbed leading to high surface temperature.
points on the green and reference roof plots. The eighteen thermocou­
During the observed period the outdoor surface temperature of the green
ples were applied on the green roof to measure indoor (Ts,in GR) and
roof was below 30 ◦ C.
outdoor (Ts,out GR) green roof surface, drainage layer (Td), growing
As can be seen in Fig. 5 outdoor surface temperature of the reference
substrate layer (Tgs), vegetation layer (Tv) temperature, and local air
roof increases rapidly and reaches its peak around noon when solar ra­
temperature at 10 cm above vegetation (T10 GR). The outdoor surface
diation is strong. High temperatures on the roof cause thermal stresses in
temperature of the green roof was measured at the interface between the
materials and have a negative impact on the durability of the roof
waterproof layer and the structural layer. The reference roof was
membrane [30]. The reference roof outdoor surface is directly exposed
equipped with nine thermocouples to measure indoor (Ts,in RR) and
to solar radiation, while the energy that the green roof received needs to
outdoor (Ts,out RR) roof surface temperature, and local air temperature at
be transferred through the conductive heat flux within the green roof
10 cm above the reference roof (T10 RR).
Four meteorological parameters were monitored to determine
background weather conditions, including ambient temperature (Ta),
relative humidity (RH), incoming solar radiation (SR) and wind speed
(WS) at 150 cm above reference roof. All sensors readings were averaged
over 10 min periods, and the data were stored in the data acquisition
system. All monitoring sensors were placed at the center of the roof plots
to avoid edge and shadow effects. The technical specifications of the
sensors with measured parameters are summarized in Table 1. Cali­
bration of the sensors was done before installation.

3. Results and discussion

3.1. Thermal performance of the green roof and reference roof

The experimental results obtained during the summer season from 3

to 6 July 2020 allowed the assessment of the summer thermal behavior
of the green roof and reference roof. During the experiment, weather
conditions were representative of the summer climate in Belgrade, with Fig. 3. The ambient temperature, relative humidity and solar radiation.

Table 1
Technical specifications of the sensors.
Sensor Model Parameters Measuring range Resolution Accuracy

Thermocouple Type K Temperature − 200 C to 1300 C

◦ ◦
1 C
◦
±1 ◦ C
T and RH sensor Testo 174H Temperature and humidity − 20 to 70 ◦ C 0.1 ◦ C ±0.5 ◦ C
0–100% 0.1% ±3%
Heat Flux meter Testo Heat flux − 2000 to 2000 W/m2 0.1 W/m2 ±2%
Pyranometer Hukseflux LP02 Solar radiation 0–2000 W/m2 1 W/m2 ±10 W/m2
Anemometer Smarty planet SPR100-01 Wind speed 0–30 m/s 0.1 m/s ±0.5 m/s
Data acquisition system Keithley – – – –
DAQ 6510

4
D. Kostadinović et al. Building and Environment 217 (2022) 109061

Fig. 4. Wind speed measured above reference roof.

Fig. 6. The indoor surface temperature of the green roof and reference roof.
(For interpretation of the references to color in this figure legend, the reader is
referred to the Web version of this article.)

roof is shown in Fig. 7. The average temperature difference of air tem­

perature above the green and reference roofs is 1.6 ◦ C in favor of the
green roof. During the observed period, the local air temperature dif­
ference between the green roof and the reference roof reaches a
maximum value of 8.6 ◦ C. It can be observed that at night little differ­
ence could be seen between these two temperatures, while the highest
difference was when solar radiation is strong. Similar results were
observed on an extensive green roof with soil substrate in China [32].
From 12:00 p.m. to 4:00 p.m. the high solar radiation and ambient
temperature cause a decrease in relative humidity and enhances the
evapotranspiration process.
With the aim of a complete understanding of green roof thermal
behavior, the vertical temperature profile through the green roof is
shown in Fig. 8. The temperature profile during sunny summer days
Fig. 5. The outdoor surface temperature of green roof and reference roof
shows that green roof reduced the daily temperature fluctuations within
structural layer. (For interpretation of the references to color in this figure the roofing system and delay heat transfer. The widest daily fluctuations
legend, the reader is referred to the Web version of this article.) and the earliest appearance of extreme values were noticed for vegeta­
tion mat temperature. Three phases can be observed in the temperature
layers. This explains that there was an average time lag of 2 h and 30 min profile through the green roof. The first phase from 00:00 a.m. to 10:00
between peaks of the outdoor surface temperature of reference and a.m. and the third phase from 6:00 p.m. to 00:00 a.m. with the same
green roof. vertical temperature sequence: outdoor roof surface > drainage layer >
During the nights from 4:00 a.m. to 6:00 a.m. the outdoor surface growing substrate layer > vegetation layer, indicating upward heat loss
temperature of the reference roof is lower than the outdoor surface from the substrate to vegetation surface. The second phase was from
temperature of the green roof with an average temperature difference of
0.2 ◦ C. The reason for this is direct exposure of the reference roof surface
to the night sky and long-wave radiative heat exchange with the sky
which permits roof cooling, while the green roof surface is covered with
plants that retain long-wave radiation under the canopy [31].
The daily temperature fluctuations of the outdoor surface of the
green roof were significantly lower compared to the reference roof and
ambient temperature fluctuations since the roof is covered with vege­
tation. The average temperature amplitude was 31.3 ◦ C for the reference
roof while 4.1 ◦ C for the green roof, which corresponds to a reduction in
temperature amplitude by 83%, reflecting the excellent insulation effect
of the green roof system.
Fig. 6 presents the indoor surface temperature of the green and
reference roof. Both roofs show similar surface temperature trends
during day and night. The daily temperature fluctuations on the indoor
surface of green roof are dampened compared to the reference roof. By
adding green roof layers, the insulation capacity of the roof system is
increased, which makes the indoor green roof surface temperature less
sensitive to changes in the ambient temperature. This is the advantage of
the green roof system in the case of summer temperature extremes. Fig. 7. The local air temperature at 10 cm above the green roof and reference
Local air temperature at 10 cm above the green roof and reference roof. (For interpretation of the references to color in this figure legend, the
reader is referred to the Web version of this article.)

5
D. Kostadinović et al. Building and Environment 217 (2022) 109061

Fig. 9. The heat flux through the green roof and reference roof in relation to
Fig. 8. The vertical temperature profile through the green roof. (For interpre­ solar radiation. (For interpretation of the references to color in this figure
tation of the references to color in this figure legend, the reader is referred to legend, the reader is referred to the Web version of this article.)
the Web version of this article.)
follows the temperature trend within the roofing systems. The green roof
10:00 a.m. to 6:00 p.m. during which the vertical temperature sequence experienced negative heat flow from 1:00 a.m. to 00:30 p.m., while
was: vegetation layer > growing substrate layer > drainage layer > positive heat flow during the remaining time. The reference roof heat
outdoor roof surface, indicating solar heat gain and thermal damping. flux was almost always positive indicating that heat was transferred
The vegetation surface has the highest averaged temperature of from the roof structure to the interior due to the strong solar radiation
25.8 ◦ C, while the average temperature of the growing substrate was and high outdoor roof surface temperature. In general, during the
25.3 ◦ C. According to Ref. [33], 27% of the solar radiation is reflected, observed period more heat is transferred into the building than trans­
60% is absorbed by the vegetation and 13% is transmitted to the roof ferred out of the building through both roofs.
structure. During the day from 10:00 p.m. to 4:00 p.m., the temperature The obtained results show lower heat flux through the green roof
of vegetation layer surface is higher than the ambient temperature, both day and night compared to the reference roof. The maximum heat
suggesting that the vegetation mat is effective in absorb and retaining flux was 7.6 W/m2 for the reference roof and 3.3 W/m2 for the green
heat. This can be explained by the following reasons: (1) the dark color roof. The green roof significantly decreased heat flow through the roof
and lower albedo value of vegetation contribute to heat absorption and by 57% and delayed heat transfer for 3 h and 40 min due to increased
incur higher temperature of vegetation layer surface; (2) non-uniform thermal mass, shading effect of vegetation, and evapotranspiration. This
heights of sedum plants and different surface roughness trap heat and is in agreement with the study [35] which compared a low insulated flat
retard the ventilation of the air layer closest to the surface. The vege­ roof to green roofs and calculated a potential reduction of 60% of heat
tation surface cooled rapidly and has a minimum temperature at night. flux during the summer period. The temperature in rooms below green
The Sedum tissues had restricted temperature buffering capacity, tend­ and reference roof was kept constant; therefore any heat flow between a
ing toward a high daytime maximum and low nighttime minimum [34]. building and its environment creates energy demand for space condi­
During day time temperature of rock mineral wool substrate was tioning. The green roofs could reduce the power consumption of the air
lower than the ambient temperature. The vegetation layer acts as a conditioning system while maintaining the same indoor air temperature
thermal insulator and significantly affects the mineral wool tempera­ [36].
ture. Shading by vegetation layer shields mineral wool substrate from During the night, heat flux through the reference roof reached a
incoming solar radiation. The outdoor roof surface temperature of the minimal value of − 0.08 W/m2 when the temperature of the outdoor
green roof reduced significantly because of the presence of vegetation, reference roof surface drops to a minimum daily value. In the early
growing substrate, and drainage layer. Substrate and drainage layer morning, the minimal value of − 3.4 W/m2 was recorded for heat flux
temperatures have compressed daily regime and delayed peaks through the green roof.
compared to the ambient temperature. There is an evident correlation between the heat flux through the
The delay in the temperature pick of substrate and drainage layer reference roof and solar radiation with a small-time lag (50 min) for the
concerning ambient temperature is respectively 40 min and 1h and 10 daytime heat flux peak, indicating the low thermal inertia of the refer­
min. Since the water reservoirs of the drainage layer are most of the time ence roof. The experimental measurements shown that main effect of
filled with air which has a low thermal conductivity, the plastic drainage green roofs was to reduce daily fluctuations of building roof temperature
layer has a good thermal insulation effect. The highest fluctuation in and correspondingly to decrease fluctuations of conductive heat flux
daily temperatures of green roof layers was noted for the vegetation through building roof. The daily fluctuations of heat flux in and out of
layer 18.2 ◦ C and the lowest for the drainage layer 8.2 ◦ C. The fluctu­ the building were lower for the green roof than for the reference roof.
ation in the daily temperature of the mineral wool substrate is 12.2 ◦ C. In The average heat flux amplitude for the green roof was 5.1 W/m2 while
terms of temperature fluctuations, the outdoor surface temperature of 6.3 W/m2 for the reference roof. The experimental measurements
the reference roof is more strongly affected by the solar radiation indicate that the lightweight green roof have improved thermal per­
compared to those measured on the outdoor surface of the green roof formance of the roof with respect to reduced conductive heat gain.
and mineral wool substrate surface.
The values of solar radiation and heat flux measured at the interface
between the indoor surface of green and reference roof and indoor air 3.2. Correlation analysis
are shown in Fig. 9. The heat flux sensors were calibrated such that a
positive reading shows heat flow into the building, while a negative Meteorological parameters play an important role in the proper
reading represents heat flow outside of the building. The heat flux evaluation of the thermal insulation and evapotranspiration effects of
the green roof. The Pearson correlation analyses were performed to

6
D. Kostadinović et al. Building and Environment 217 (2022) 109061

evaluate the influence of the local meteorological parameters, including Table 3

ambient temperature (Ta), relative humidity (RH), solar radiation (SR), Pearson correlation coefficients between meteorological parameters and green
and wind speed (WS) on thermal performance parameters of both roofs. roof cooling effect.
In addition, correlation coefficients of meteorological parameters and ΔTs,out ΔT10 Δq
cooling effect of green roof reflected by green-reference roof differences Ta − 0.735* − 0.705* − 0.658*
(ΔTs,out, ΔT10, and Δq) were determined. RH 0.601* 0.676* 0.594*
SR − 0.790* − 0.565* − 0.691*
3.2.1. Effect of meteorological parameters on thermal performance of the WS − 0.230* − 0.156* − 0.154*
green roof and reference roof * Correlation is significant at the 0.001 level (1-tailed).
The Pearson correlation coefficients between meteorological pa­
rameters and thermal performance of the green and reference roof are prevalent role in regulating the green roof cooling effect. SR has the
presented in Table 2. Generally, the results demonstrate that the thermal highest influence on outdoor roof surface temperature differences,
performance of the reference roof was more sensitive to meteorological suggesting that a better cooling effect can be achieved on hot days with
parameters than the green roof. It was found that Ta and RH have the high SR. High SR on the green roof can increase latent heat loss by
strongest correlation coefficients, indicating their dominant role on the enhancing evapotranspiration. The relationships between RH and the
thermal performance of the green and reference roof. green roof cooling effect were opposite those of the SR and Ta. High RH
In contrast to the green roof, heat flux through the reference roof was is associated with low SR and corresponds to lower air and surface
highly correlated with SR, indicating the temperature buffering of the temperature and restricted temperature reduction. The results show
green roof layers under solar radiation. The correlation of SR with heat little influence from WS on the observed temperature and flux differ­
flux through the reference roof was positive with a high value, implying ences. Negative low correlation coefficients were found for WS, showing
that the stronger SR was the more heat flow into the room below the that wind can enhance the evapotranspiration cooling of the green roof.
reference roof. For the green roof, correlations of meteorological pa­ The difference of outdoor roof surface temperature has the highest
rameters were stronger with the temperature of the vegetation layer but correlation coefficients and the sequence of meteorological parameters
weaker with the temperature of the middle layers, such as the substrate is solar radiation > ambient temperature > outdoor humidity > wind
layer and the outdoor surface of the structural layer. speed. It can be concluded that better passive cooling effect of green roof
It can be noted that vegetation surface temperature is most corre­ can be achieved on hot days with high SR.
lated with the meteorological parameters and the sequence of meteo­
rological parameters is ambient temperature > solar radiation > relative 3.3. Comparison with other conventional green roof systems
humidity > wind speed. High SR caused more heat absorbed by the
plants, which generate high vegetation surface temperature. For sub­ Compared with findings of the study in the same climatic zone in
strate temperature, the sequence of meteorological factors is a little Italy [39], the green roof with mineral wool substrate achieves a much
different compared to vegetation temperature: ambient temperature > higher reduction of heat flux through roof. This is probably because of
relative humidity > solar radiation > wind speed. The thermal effect of the low transmittance of the roof systems in Italian study and the fact
SR on the mineral wool substrate layer is less compared to that on the that the building did not have air conditioning.
vegetation layer due to the shading effect of vegetation. The measurements conducted during summer in NYC [29] showed
SR and Ta were both positively correlated with all temperatures and that modular green roof with soil substrate reduces the outdoor roof
heat flux on the green and reference roof, while RH is just the opposite surface temperature by 33.9 ◦ C. Reduction in maximum outdoor roof
because of the relation between high RH and low SR and low Ta. The surface temperature on sunny summer days in our study is lower by 6 ◦ C.
high RH combined with low Ta suppresses evapotranspiration (as the This could be assigned to the high value of the black membrane roof of
water potential gradient is reduced) hence weakened the green roof 80 ◦ C, compared to our reference roof.
thermal performance [37]. Wind can reduce surface and local air tem­ In this study cooling of local air by vegetation was the most pro­
perature by increasing green roof evapotranspiration [38]. Weak cor­ nounced during the hottest time of the day. Another study on the ther­
relation coefficients with consistently positive signs were observed for mal performance of tropical green roof has been conducted in humid
WS. The reason could be heat advection from surroundings. tropical Hong Kong [40]. The study concluded that the green roof with
soil substrate and rockwool water storage has the strongest cooling ef­
3.2.2. Effect of meteorological parameters on green roof cooling effect fect during the day when high solar radiation can intensify shading and
The cooling effect of a green roof can be considered by three items: evaporation effects.
ΔTs,out, ΔT10, and Δq, which are the differences of outdoor roof surface The vegetation surface temperature has the highest averaged tem­
temperature, local outdoor air temperature, and heat flux between a perature and highest fluctuation in daily temperatures, suggesting that
green roof and reference roof. The Pearson correlation coefficients be­ the vegetation mat is effective in retaining heat. This is consistent with
tween meteorological parameters and the green roof cooling effect are other field studies [40)], [41].
reported in Table 3. The temperature of the mineral wool substrate and drainage layer
SR and Ta have the strongest correlation coefficients, reflecting their was higher during the night and morning hours as compared to the
ambient temperature. The study from Germany [42] shows that sub­
strate which consists of crushed bricks, clayey slate, compost, and peat
Table 2
Pearson correlation coefficients between meteorological parameters and ther­ retained heat during the night such that these temperatures were higher
mal performance of the green roof and reference roof. than the ambient air temperatures.
Most results from this study are in accordance with the literature
Green roof Reference roof
regard to conventional green roof systems and can justify green roofs
Ts,out Tgs Tv q Ts,out q with rock mineral wool substrate as a good passive cooling technique.
Ta 0.674* 0.922* 0.940* 0.576* 0.816* 0.961* This type of lightweight green roof significantly contributes to building
RH − 0.534* − 0.767* − 0.795* − 0.529* − 0.682* − 0.908* insulation and provides cooling during hot summer days.
SR 0.394* 0.766* 0.891* 0.229* 0.832* 0.809*
WS 0.298* 0.279* 0.269* 0.126** 0.273* 0.229*

* Correlation is significant at the 0.001 level (1-tailed).

** Correlation is significant at the 0.05 level (1-tailed).

7
D. Kostadinović et al. Building and Environment 217 (2022) 109061

4. Conclusions interests or personal relationships that could have appeared to influence

the work reported in this paper.
This study investigates the summer thermal performance of a light­
weight extensive green roof system with rock mineral wool substrate Acknowledgment
installed on a school building. In-situ measurements during a summer
period included heat fluxes through the lightweight green and reference The research was funded by the Ministry of Education, Science and
roof, the vertical temperature profile of both roofs, local outdoor air Technological Development of the Republic of Serbia. Special thanks to
temperature above roofs, and local meteorological parameters. Aleksandar Milićević for the paper review and Tamara Petrović for
A very small number of experimental researches are conducted on language editing.
the thermal behavior of a lightweight green roof in different climatic
zones. In order to better understand green roof summer thermal References
behavior in a humid subtropical climate, correlations between meteo­
rological parameters and thermal performance of the lightweight green [1] D.B. Rowe, Green roofs as a means of pollution abatement, Environ. Pollut. 159
(8–9) (2011) 2100–2110, https://doi.org/10.1016/j.envpol.2010.10.029.
roof and reference roof and the green roof cooling effect are analysed. [2] J. Mentens, D. Raes, M. Hermy, Green roofs as a tool for solving the rainwater
The main findings are summarized as follows: runoff problem in the urbanized 21st century? Landsc. Urban Plann. 77 (3) (2006)
217–226, https://doi.org/10.1016/j.landurbplan.2005.02.010.
[3] H. Takebayashi, M. Moriyama, Surface heat budget on green roof and high
• The green roof reduced the maximum outdoor roof surface temper­ reflection roof for mitigation of urban heat island, Build. Environ. 42 (8) (2007)
ature by 27.5 ◦ C and daily temperature fluctuations experienced by 2971–2979, https://doi.org/10.1016/j.buildenv.2006.06.017.
the roof by 83%, reflecting the excellent insulation effect of the green [4] T. Van Renterghem, D. Botteldooren, Reducing the acoustical façade load from
road traffic with green roofs, Build. Environ. 44 (5) (2009) 1081–1087, https://doi.
roof system with a mineral wool substrate. org/10.1016/j.buildenv.2008.07.013.
• The green roofs reduce the influence of the ambient temperature on [5] N.S.G. Williams, J. Lundholm, J.S. MacIvor, FORUM: do green roofs help urban
the indoor green roof surface temperature. biodiversity conservation? J. Appl. Ecol. 51 (6) (2014) 1643–1649, https://doi.
org/10.1111/1365-2664.12333.
• The average air temperature decrease was 1.6 ◦ C, indicating a good
[6] B. Raji, M.J. Tenpierik, A. van den Dobbelsteen, The impact of greening systems on
cooling effect of a lightweight green roof at 10 cm height compared building energy performance: a literature review, Renew. Sustain. Energy Rev. 45
to an ad adjacent concrete roof. (2015) 610–623, https://doi.org/10.1016/j.rser.2015.02.011.
• The vertical temperature profile shows that the green roof reduced [7] C.R. Friedrich, Principles for selecting the proper components for a green roof
growing media, in: Proceedings of 3rd North American Green Roof Conference:
the daily temperature fluctuations within the roofing system and Greening Rooftops for Sustainable Communities, 2005, pp. 262–274. Washington
delay heat transfer, indicating that the green roofs with mineral wool DC.
substrate can be used in passive building design as an efficient [8] B. Garner, M. Furbish, W. Griffin, L.L.C. Furbish Company, Mineral wool in green
roofs, Available at: https://furbishco.com/wp-content/uploads/2015/06/Mineral-
thermal barrier. Wool-in-Green-Roofs.pdf. (Accessed 28 September 2021).
• The green roof with mineral wool substrate can serve as a thermal [9] P. Bevilacqua, D. Mazzeo, R. Bruno, N. Arcuri, Experimental investigation of the
mass, reducing heat flow through the roof by 57%. thermal performances of an extensive green roof in the Mediterranean area, Energy
Build. 122 (2016) 63–79, https://doi.org/10.1016/j.enbuild.2016.03.062.
• The ambient temperature and relative humidity are the most corre­ [10] M. Maiolo, B. Pirouz, R. Bruno, S.A. Palermo, N. Arcuri, P. Piro, The role of the
lated factors, indicating their dominant role in the thermal perfor­ extensive green roofs on decreasing building energy consumption in the
mance of green and reference roofs. Heat flux through the reference mediterranean climate, Sustainability 12 (1) (2020) 359, https://doi.org/10.3390/
su12010359.
roof was highly correlated with SR, indicating the temperature [11] M. Foustalieraki, M.N. Assimakopoulos, M. Santamouris, H. Pangalou, Energy
buffering of the green roof layers under solar radiation. performance of a medium scale green roof system installed on a commercial
• The green roof cooling effect is most pronounced in conditions of building using numerical and experimental data recorded during the cold period of
the year, Energy Build. 135 (2017) 33–38, https://doi.org/10.1016/j.
high solar radiation and low relative humidity.
enbuild.2016.10.056.
[12] C.Y. Jim C, Heat-sink effect and indoor warming imposed by tropical extensive
The results from this study demonstrate that green roofs with rock green roof, Ecol. Eng. 62 (2014) 1–12, https://doi.org/10.1016/j.
mineral wool substrate can be suitable urban greening solutions since ecoleng.2013.10.022.
[13] M. Santamouris, C. Pavlou, P. Doukas, G. Mihalakakou, A. Synnefa, A. Hatzibiros,
they contribute to building insulation and provide cooling during sum­ P. Patargias, Investigating and analyzing the energy and environmental
mer while provides an adequate plant growth media. This paper can performance of an experimental green roof system installed in a nursery school
help researchers and practitioners to have a better understanding of the building in Athens, Greece, Energy 32 (9) (2007) 1781–1788, https://doi.org/
10.1016/j.energy.2006.11.011.
thermal performance and cooling effect of lightweight green roof in [14] L.F.M. Francis, M.B. Jensen, Benefits of green roofs: a systematic review of the
correlation with meteorological parameters in a humid subtropical evidence for three ecosystem services, Urban For. Urban Forestry and Urban
climate. The obtained results could be applied to encourage the instal­ Greening 28 (2017) 167–176, https://doi.org/10.1016/j.ufug.2017.10.015.
[15] C. Jim, S. Tsang, Biophysical properties and thermal performance of an intensive
lation of green roof systems on public and school buildings. green roof, Build. Environ. 46 (6) (2011) 1263–1274, https://doi.org/10.1016/j.
Further analysis must be done so that the benefits of the lightweight buildenv.2010.12.013.
green roof can be more fully evaluated. This should include analysis of [16] J.T. Lundholm, B.M. Weddle, J.S. MacIvor, Snow depth and vegetation type affect
green roof thermal performance in winter, Energy Build. 84 (2014) 299–307,
winter thermal behavior of the lightweight green roof and influence of https://doi.org/10.1016/j.enbuild.2014.07.093.
water content in mineral wool substrate on thermal performance of [17] T.G. Theodosiou, Summer period analysis of the performance of a planted roof as a
green roof. passive cooling technique, Energy Build. 35 (9) (2003) 909–917, https://doi.org/
10.1016/S0378-7788(03)00023-9.
[18] N.H. Wong, D.K.W. Cheong, H. Yan, J. Soh, C.L. Ong, A. Sia, The effects of rooftop
CRediT authorship contribution statement garden on energy consumption of a commercial building in Singapore, Energy
Build. 35 (4) (2003) 353–364, https://doi.org/10.1016/S0378-7788(02)00108-1.
Danka Kostadinović: Writing – original draft, Methodology, [19] C. Feng, Q. Meng, Y. Zhang, Theoretical and experimental analysis of the energy
balance of extensive green roofs, Energy Build. 42 (6) (2010) 959–965, https://doi.
Investigation, Conceptualization. Marina Jovanović: Writing – review org/10.1016/j.enbuild.2009.12.014.
& editing, Supervision, Investigation. Vukman Bakić: Writing – review [20] A. Nardini, S. Andri, M. Crasso, Influence of substrate depth and vegetation type on
& editing, Supervision, Investigation. Nenad Stepanić: Supervision, temperature and water runoff mitigation by extensive green roofs: shrubs versus
herbaceous plants, Urban Ecosyst. 15 (2012) 697–708, https://doi.org/10.1007/
Investigation. Maja Todorović: Supervision. s11252-011-0220-5.
[21] Ciril Arkar, Suzana Domjan, Sašo Medved, Heat transfer in a lightweight extensive
Declaration of competing interest green roof under water-freezing conditions, Energy Build. 167 (2018) 187–199,
https://doi.org/10.1016/j.enbuild.2018.02.056.

The authors declare that they have no known competing financial

8
D. Kostadinović et al. Building and Environment 217 (2022) 109061

[22] B.-S. Lin, C.-C. Yu, A.-T. Su, Y.-J. Lin, Impact of climatic conditions on the thermal symposium on heating, ventilation and air conditioning (ISHVAC) and the 3rd
effectiveness of an extensive green roof, Build. Environ. 67 (2013) 26–33, https:// international conference on building energy and environment (COBEE), Procedia
doi.org/10.1016/j.buildenv.2013.04.026. Eng. 121 (2015) 15971604, https://doi.org/10.1016/j.proeng.2015.09.186.
[23] S.W. Tsang, C.Y. Jim, Theoretical evaluation of thermal and energy performance of [33] P.M. Krusche, D. Althaus, I. Gabriel, M. Weig-Krusche, Ökologisches Bauen
tropical green roofs, Energy 36 (5) (2011) 3590–3598, https://doi.org/10.1016/j. Herausgegeben Vom Umweltbundesamt, 1982, https://doi.org/10.1007/978-3-
energy.2011.03.072. 322-84893-2. Bauverlag GmbH, Wiesbaden & Berlin.
[24] Y. He, H. Yu, N. Dong, H. Ye, Thermal and energy performance assessment of [34] C.Y. Jim, Assessing climate-adaptation effect of extensive tropical green roofs in
extensive green roof in summer: a case study of a lightweight building in Shanghai, cities, Landsc. Urban Plann. 138 (2015) 54–70, https://doi.org/10.1016/j.
Energy Build. 127 (2016) 762–773, https://doi.org/10.1016/j. landurbplan.2015.02.014.
enbuild.2016.06.016. [35] N.H. Wong, Y. Chen, C.L. Ong, A. Sia, Investigation of thermal benefits of rooftop
[25] L.L.H. Peng, X. Yang, Y. He, Z. Hu, T. Xu, Z. Jiang, L. Yao, Thermal and energy garden in the tropical environment, Build. Environ. 38 (2003) 261–270, https://
performance of two distinct green roofs: temporal pattern and underlying factors in doi.org/10.1016/S0360-1323(02)00066-5.
a subtropical climate, Energy Build. 185 (2019) 247–258, https://doi.org/ [36] W. Yang, Z. Wang, J. Cui, Z. Zhu, X. Zhao, Comparative study of the thermal
10.1016/j.enbuild.2018.12.040. performance of the novel green (planting) roofs against other existing roofs,
[26] D. Dimitrijević Jovanović, D. Kostadinović, P. Živković, D. Ranđelović, Green Sustain. Cities Soc. 16 (2015) 1–12, https://doi.org/10.1016/j.scs.2015.01.002.
living roofs as a part of green infrastructure, in: Proceedings of 3rd International [37] M. Santamouris, Cooling the cities - a review of reflective and green roof mitigation
Conference on Urban Planning - ICUP2020, 2020, pp. 171–178, 12-13 November. technologies to fight heat island and improve comfort in urban environments, Sol.
[27] M. Kottek, J. Grieser, C. Beck, B. Rudolf, F. Rubel, World map of the Köppen-Geiger Energy 103 (2014) 682–703, https://doi.org/10.1016/j.solener.2012.07.003.
climate classification updated, Meteorol. Z. 15 (3) (2006) 259–263, https://doi. [38] C.Y. Jim, L.L.H. Peng, Weather effect on thermal and energy performance of an
org/10.1127/0941-2948/2006/0130. extensive tropical green roof, Urban For. Urban Green. 11 (2012) 73–85, https://
[28] ISO 6946:2017-Building Components and Building Elements-Thermal Resistance doi.org/10.1016/j.ufug.2011.10.001.
and Thermal Transmittance-Calculation Methods, International Organization for [39] P. Bevilacqua, The effectiveness of green roofs in reducing building energy
Standardization: (Geneva, Switzerland). consumptions across different climates, A summary of literature results, Renew.
[29] S.R. Gaffin, C. Rosenzweig, J. Eichenbaum-Pikser, R. Khanbilvardi, T. Susca, Sustain. Energy Rev. 151 (2021) 111523, https://doi.org/10.1016/j.
A Temperature and Seasonal Energy Analysis of Green, White, and Black Roofs, rser.2021.111523.
Columbia University, Center for Climate Systems Research, New York, 2010. [40] C.Y. Jim, Effect of vegetation biomass structure on thermal performance of tropical
Technical Report. green roof, Landsc. Ecol. Eng. 8 (2012) 173–187, https://doi.org/10.1007/s11355-
[30] B. Schindlera, L. Blausteina, A. Vasla, G. Kadasb, M. Seifanc, Cooling effect of 011-0161-4.
Sedum sediforme and annual plants on green roofs in a Mediterranean climate, [41] M. Eksi, D.B. Rowe, I.S. Wichman, J.A. Andresen, Effect of substrate depth,
Urban For. Urban Green. 38 (2019) 392–396, https://doi.org/10.1016/j. vegetation type, and season on green roof thermal properties, Energy Build. 145
ufug.2019.01.020. (2017) 174–187, https://doi.org/10.1016/j.enbuild.2017.04.017.
[31] A. Solcerova, F. van de Ven, M. Wang, M. Rijsdijk, N. van de Giesen, Do green roofs [42] J. Franzaringa, L. Steffana, W. Ansel, R. Walker, A. Fangmeier, Water retention,
cool the air? Build. Environ. 111 (2017) 249–255, https://doi.org/10.1016/j. wash-out, substrate and surface temperatures of extensive green roof mesocosms-
buildenv.2016.10.021. Results from a two year study in SW-Germany, Ecol. Eng. 94 (2016) 503–515,
[32] Y. Hea, H. Yua, M. Zhaoa, Thermal performance study of extensive green roof in https://doi.org/10.1016/j.ecoleng.2016.06.021.
Shanghai district: a case study of lightweight building in winter, 9th international