Applied Energy 107 (2013) 10–24

Contents lists available at SciVerse ScienceDirect

Applied Energy
journal homepage: www.elsevier.com/locate/apenergy

Evaluation of Overall Thermal Transfer Value (OTTV) for commercial

buildings constructed with green roof
A.L.S. Chan ⇑, T.T. Chow
Division of Building Science and Technology, College of Science and Engineering, City University of Hong Kong, Tat Chee Avenue, Hong Kong

h i g h l i g h t s

" Correction factors ranging from 0.03 to 0.99 were derived for OTTV evaluation of building with green roof.
" An experimental setup of green roof was constructed on the rooftop of a building.
" Field measurement was conducted in the green roof system.
" Computer simulation model (EnergyPlus with Ecoroof) was validated using experimental data of the green roof.

a r t i c l e i n f o a b s t r a c t

Article history: Overall Thermal Transfer Value (OTTV) is a measure of average heat gain into a building through the
Received 19 September 2012 building envelope. It is a widely adopted measure in many countries for enhancing energy-efficient build-
Received in revised form 12 January 2013 ing design. In the past decade, there is increasing application of green roof into commercial buildings for
Accepted 2 February 2013
enhanced building insulation, leading to reduction in heat gain through the roof area as well as cooling
Available online 6 March 2013
requirement of a building. Since the current OTTV equations and coefficients were originally developed
for buildings with traditional bare roof construction, building designers have difficulty to compute the
Keywords:
OTTV for building constructed with green roof. The aim of this study is to revise the existing OTTV calcu-
Overall Thermal Transfer Value (OTTV)
Green roof system
lation method and derive a set of correction factors for OTTV evaluation of green roof integrated buildings.
EnergyPlus An experimental setup of a green roof system with sensors was installed on the rooftop of a commercial
Correction factors building. The measured data were used for validation of a building energy simulation program Energy-
Building energy simulations Plus incorporated with a green roof model Ecoroof. Four building cases with typical and traditional roof
constructions were modeled using the validated computer simulation program. Through a series of para-
metric computer simulations, a correlation between OTTV and annual heat gain through the roof area was
established with that a set of correction factors ranging from 0.03 to 0.99 was developed. Theses correc-
tion factors can be used by building designers to compute the OTTV of building constructed with green
roof. The details of methodology and findings are reported in this paper.
Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction adopted in the building industry for governing energy-efficient

building/building services system design. In Hong Kong, five pre-
In recent years, there is cooperative effort to reduce greenhouse scriptive energy codes have been published by the local govern-
gas (GHG) emission all over the world. The main objective is to ment as listed below [1–5]:
pursue a blue sky and healthy environment for sustainable devel-
opment of a society. Minimization of energy use is one of the keys i. Code of Practice for Energy Efficiency of Lighting
to foster GHG reduction. Building, as one of the largest electricity Installations.
consumers, can make a marked contribution to energy conserva- ii. Code of Practice for Energy Efficiency of Air Conditioning
tion as well as GHG removal by sophisticated building design. Installations.
Currently there are two major types of building energy codes, iii. Code of Practice for Energy Efficiency of Electrical
namely prescriptive and performance-based energy codes, widely Installations.
iv. Code of Practice for Energy Efficiency of Lift & Escalator
Installations.
v. Code of Practice for Overall Thermal Transfer Value in
⇑ Corresponding author. Tel.: +852 3442 9676; fax: +852 3442 9716.
Buildings.
E-mail address: [email protected] (A.L.S. Chan).

0306-2619/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.apenergy.2013.02.010
 A.L.S. Chan, T.T. Chow / Applied Energy 107 (2013) 10–24 11

Nomenclature

Af area of fenestration (m2) Xsim,i surface temperature (°C) at or heat flux (W/m2) through
Ar area of opaque roof (m2) a bare roof/green roof at hour i simulated by EnergyPlus
Aw area of opaque wall (m2) z height or depth (m)
CF correction factor
ESM external shading multiplier Greeks
Hf foliage sensible heat flux (W/m2) DTwin temperature difference for window glass (°C)
Hg ground sensible heat flux (W/m2) af albedo (short-wave reflectivity) of the canopy
I#ir total incoming long-wave radiation (W/m2) ag albedo (short-wave reflectivity) of ground surface
I#s total incoming short-wave radiation (W/m2) ar absorptivity of opaque roof
Lf foliage latent heat flux (W/m2) aw absorptivity of opaque wall
Lg ground latent heat flux (W/m2) ef emissivity of canopy
OTTVi,BR OTTV of a bare roof of a building case i eg emissivity of the ground surface
SC shading coefficient of fenestration e1 eg + e f  eg ef
SF solar factor for vertical surface (W/m2) j soil thermal conductivity of the surface (W/mK)
Tf foliage temperature (K) r Stefan–Boltzmann constant
Tg ground surface temperature (K) rf fractional vegetation coverage
TDEQr equivalent temperature difference for opaque roof (°C)
TDEQw equivalent temperature difference for opaque wall (°C) Subscripts
Ur thermal transmittance of opaque roof (W/m2 °C) i building case A, B, C or D
Uw thermal transmittance of opaque wall (W/m2 °C) j soil thickness (m)
Xexp,i surface temperature (°C) at or heat flux (W/m2) through k height of plant (m)
a bare roof/green roof at hour i measured from experi- l Leaf Area Index (LAI)
ment m calendar month (from April to October)

Prescriptive building energy codes are simple and can provide a In 2004, an OTTV-based energy estimation model for commer-
straightforward approach for building designers to evaluate the cial buildings in Thailand was developed by Chirarattananon and
compliance of a building/building services system design with Taveekun [14]. Simulation program DOE-2 was utilized to conduct
the energy codes. a series of parametric runs to develop OTTV formulations for four
On the other hand, performance-based building energy code is different types of commercial buildings in Thailand. The resulting
an alternative path to the prescriptive codes [6]. It considers the OTTVs were used in further parametric runs to develop a formula-
various components of building energy consumption, allowing tion for the cooling coil load and energy use of the commercial
trade-off among them. This approach provides rooms to building buildings. The results were expected to have contribution towards
designers for innovative design. It focuses on the total energy con- energy code compliance and energy monitoring.
sumption of a building design which is termed as Design Energy. A Kunchornrat et al. proposed new parameters for OTTV calcula-
corresponding reference building (having the same size and shape tion corresponding to the real climate zones in Thailand in 2009
as the design building) that fully complies with all the prescriptive [15]. The impacts of TDEQw and DTwin were examined and proposed
building energy codes can be developed and its total energy con- to the existing OTTV calculation. The study found that the new
sumption is calculated as Energy Budget. The performance-based parameters could influence the OTTV of a reference building, which
building energy code is deemed to be complied if the Design varied from 2.4% to 9.1% within different climate zones.
Energy is smaller than or equal to the Energy Budget. No matter Chua and Chou refined the original OTTV equation adopted in
which approach is adopted, the Overall Thermal Transfer Value Singapore and applied into residential buildings in 2010 [16]. They
(OTTV) is a basic requirement that a building design must fulfill. found that the original OTTV equation did not accurately account
for the relative components of heat gains through the building
1.1. Development of OTTV control envelope. By using Singapore’s weather data consolidated for a
particular year, three coefficients (TDEQw, DTwin and SF) were de-
The use of OTTV was firstly proposed by the American Society of rived by performing several multi-parametric simulations on two
Heating, Refrigerating and Air-conditioning Engineers (ASHRAE) in residential building types. An Envelope Thermal Transfer Value
1975 [7]. It is a measure of the average heat gain into a building (ETTV) equation was then developed for the residential buildings
through the building envelope and can be used for comparing in Singapore.
the thermal performance of buildings. There are three major com- Research on OTTV in Hong Kong had been carried out by various
ponents involved in the OTTV calculation: (i) conduction through local researchers. Chow & Chan had used DOE-2 program to
opaque wall, (ii) conduction through fenestration, and (iii) solar conduct parametric studies to determine OTTV equations and
radiation through fenestration. It is usual to have two sets of OTTV coefficients for building envelopes in Hong Kong [17]. The
for a building, namely OTTV for external walls (OTTVwall) and OTTV window-to-wall ratio and orientation of building were varied in
for roof (OTTVroof). The OTTV of an entire building is given by the the study. They argued that heat transmission through the building
weighted average of the OTTVs of the external walls and the roof envelope might reverse in direction during certain air-conditioned
of a building. hours in a year. Therefore, two values of OTTV, one for the summer
In Asia, Singapore was the first country to develop an OTTV stan- and the other for the winter season, had been established to
dard and regulatory control over the OTTV of air-conditioned build- account for seasonal changes in Hong Kong. The summer OTTV,
ings since 1979 [8,9]. Some other Asian countries including calculated from the heat gain in hot season, was recommended
Philippines, Thailand, Malaysia and Indonesia also developed their as more appropriate for evaluating the thermal performance of
OTTV standards in 1980s and 1990s [10–13]. building envelope in Hong Kong.
12 A.L.S. Chan, T.T. Chow / Applied Energy 107 (2013) 10–24

Hui had reviewed the concept and history of OTTV methods ance on thermal reduction and planting arrangement for green
[18]. Limitations of the OTTV standard were discussed. It was re- roofs.
ported that OTTV method only dealt with the building envelope Ouldboukhitine et al. investigated experimentally on three
and did not consider other aspects of building design such as light- main physical properties of green roofs with an aim at determining
ing and air-conditioning systems. Performance-based building en- the key green roof modeling parameters [26]. Correlation between
ergy code was proposed as an alternative path to code compliance. thermal conductivity of substrate with water content, sorption and
Similar discussion was presented by Yik and Wan [19]. desorption isotherms for different temperatures and moisture buf-
Ho also had a review on OTTV in Hong Kong with regard to its fer capacity, and porosity range of the substrate in green roof sys-
legislative requirements and control [20]. Problems faced by build- tem were investigated. The objective is to estimate the parameters
ing professionals in calculation of OTTV for complicated commer- used as input data in developed green roof model to evaluate the
cial buildings were also discussed. A user-friendly computer energy performance of building constructed with green roof.
software for modeling the thermal performance of a building enve- In addition to the investigation by experiments, numerical
lope design and checking the compliance of the OTTV requirement study had been conducted by various researchers to study the per-
were introduced. formance of green roof systems. Luikov’s model [27] had been used
The OTTV standard in Hong Kong was developed from a consul- by a number of researchers. Palomo [28] studied the cooling poten-
tancy study commissioned by the Hong Kong Government in 1991 tial of green roof in summer using Luikov’s model by parametric
[21]. The consultancy study revealed that conduction heat gain sensitivity study. It was revealed that green roof, acting as an insu-
through fenestration had insignificant effect when compared to lating device, could reduce heat flux through the roof area. How-
the other two components of heat gain. As a result, the OTTV equa- ever, improving green roof for summer may causes problems in
tions adopted in the OTTV standard in Hong Kong only consists of winter time. The same model had been used by Kumar and Kaushik
conduction through opaque wall and solar radiation through [29] to evaluate the performance of green roof systems. Parametric
fenestration. simulations had been carried out using an in-house developed
In 1995, the Hong Kong Government launched an OTTV control building energy simulation code to study the variations in canopy
on the design of commercial buildings (35 W/m2 for building tower air temperature, entering heat flux through roof area and indoor air
and 80 W/m2 for podium) [5]. After a latest review in 2011, the temperature. A simplified version of Palomo’s analytical model was
OTTV limits had been tightened up to 24 W/m2 for building tower applied by Theodosiou [30] to investigate and identify the various
and 56 W/m2 for podium respectively [22]. characteristics and climatic factors that could affect the perfor-
mance of a planted roof. Foliage density (leaf area index), insula-
1.2. Research on green roof tion layer thickness and relative humidity were found as
important factors on the thermal performance of green roof sys-
In the past decade, there is significant change in the design and tems. Alexandri and Jones [31] had constructed two test cells,
construction of commercial buildings with improved thermal insu- one carrying plain pavement grey slabs and another one was a
lation and energy-efficient component such as green roof system. green test cell where a vegetated medium was placed on top of
Green roof is a vegetated space that is structurally integrated on the pavement slabs. Temperature, relative humidity, convective
the top of a building roof. It offers a number of merits such as ame- heat transfer coefficient and stomatal resistance had been mea-
nity and aesthetic benefits including increased leisure and func- sured. The results were used for validation of a one-dimensional
tional open space. In terms of environmental benefit, urban heat heat and mass transfer algorithm developed for studying the effect
island effect can be mitigated since the plants can offer the roof of green roof, based on Luikov’s model. Other researchers devel-
an increased building insulation, protect the buildings from solar oped their own models [32–34], which were validated by results
radiation and carry out heat energy through evaporation process measured from experimental setup, to study the performance of
of the plant and soil. With a green roof system, the heat gain as different green roof systems.
well as the cooling requirement of a building can be reduced. Abalo et al. [35] developed a model for evaluating the cooling
The idea of green roof is widely promoted and becomes popular potential of green roofs. Heat and mass transfer equations were
in many cities. Wong et al. [23,24] had conducted field measure- solved using a finite difference scheme and Thomas algorithm.
ment on the roofs of selected buildings in Singapore to investigate The effect of leaf area index and Biot number on the diurnal varia-
the thermal impacts of rooftop garden. It was found that rooftop tion of solar heat gain factor was analyzed. The investigation re-
garden could contribute thermal benefits to both buildings and vealed that the foliage density and the vegetable canopy type
their surrounding environment. From the measurement on the roof selection could influence the thermal efficiency of the bioclimatic
of a low-rise commercial building, some quantitative data were ob- insulation screen significantly.
tained. A maximum decrease in the surface temperature caused by Some researchers applied existing building energy simulation
the plants was around 30 °C. Less long-wave radiation emitted program to investigate the thermal performance of green roof sys-
from the planted roof was confirmed through comparisons of glo- tems. One of the computer software used is TRNSYS. Researchers
bal temperatures and mean radiant temperature (MRT). Maximum firstly conducted measurement on green roof systems to prelimi-
differences of the global temperature and MRT were 4.05 °C and narily study their performance. Then the data collected were used
4.5 °C, respectively. It gave evidence that planted roofs could effec- for development of a simple green roof model [36–39]. The predic-
tively mitigate the urban heat island effect in urban environment. tive numerical model was then incorporated in TRNSYS to investi-
Experiment in a test chamber had been carried out by Fang [25] gate the thermal and energy performances of buildings with
for investigating the thermal reduction effect of plant layers on different green roof systems, under various meteorological condi-
rooftop. A chamber of dimensions 40 cm (L)  10 cm tions. The performance evaluation in general showed a significant
(W)  70 cm (H) was constructed on the rooftop of a building in reduction of the building cooling load in summer season. This reduc-
Taiwan and covered with heat insulation. The plant was located in- tion varied in a range of 6–49% for the whole building and in a range
side the chamber with thermocouples placed to measure the soil of 12–87% for the roof floor. However, the influence of the green roof
temperature. Two relevant parameters: coverage ratio and total system in the building heating load was found insignificant.
leaf thickness were found positively correlated with the thermal DOE-2, another well-known building energy simulation pro-
reduction ratio (TRR). With the measured data, quantitative TRR gram was used by Wong et al. [40] to determine the effects of roof-
map was established which could provide a straightforward guid- top garden on the cooling load and annual energy consumption of a
 A.L.S. Chan, T.T. Chow / Applied Energy 107 (2013) 10–24 13

five-story hypothetical building in Singapore. Thermal resistances ings. A light-weight and extensive green roof technology planted
of different plants (turfing, shrubs and trees) with different soil with a drought-tolerant succulent plant called Stonecrop was iden-
types were calculated using measured data and then input into tified and tested with different drainage and substrate combina-
DOE-2 program. The results showed a saving of 0.6–14.5% in the tions with a view of finding an appropriate package of
annual energy consumption of air-conditioning system and shrubs recommendations for energy conservation and collateral benefits
was found to be most effective in reducing building energy in Hong Kong.
consumption. In 2010, He and Jim [50] developed a simulation model based a
In 2008, Sailor [41] had developed and integrated a model of en- traditional Bowen ratio energy balance model (BREBM) and a pro-
ergy and moisture balance of vegetated rooftop into a building en- posed solar radiation shield effectiveness model (SEM) to study the
ergy simulation program EnergyPlus. The model formulation was thermodynamic transmission in green roof ecosystems under dif-
based on the Army Corps of Engineers’ FASST vegetation model ferent vegetation treatments. Moreover, Jim et al. [51,52] carried
[42] which solves simultaneously for the foliage surface and soil out investigation to study the weather effect on the thermal per-
temperatures at each time step. FASST tracks the energy and mois- formance of a retrofitted extensive green roof on a railway station
ture balance within a vegetated soil. It is a one-dimensional model in humid subtropical Hong Kong. The impacts of solar radiation,
which contains energy budgets for both the foliage layer and the relative humidity, soil moisture and wind speed were explored.
ground surface. Through the study, it was found that green roof passive cooling
A preliminary validation of the model was carried out by Sailor was enhanced by high solar radiation and low relative humidity
[41] using measured soil temperatures at the University of Central in typical sunny summer days. High soil moisture supplemented
Florida test site which showed sufficient confidence that the model by irrigation could lower the air and vegetation surface tempera-
could perform adequately. Model sensitivity had been tested for ture, and dampen the diurnal temperature fluctuations. Moreover,
two cities, Chicago and Houston, with three key parameters – soil high wind speed could increase the evaportranspiration cooling of
thickness, vegetative cover (leaf area index, LAI) and irrigation. It green roof area.
was found that both soil thickness and LAI had significant impact The studies from various researchers reveal that green roof sys-
on the building energy use. However, Chicago simulations showed tem can significantly improve the thermal performance of building
more wintertime sensitivity to green roof design variations while and the installation of green roof is becoming more popular in
Houston simulations demonstrated more summertime sensitivity. modern cities. However, building designers have difficulty to com-
The response was significantly dependent on the building location pute the OTTVroof for buildings constructed with green roof since
(climate). the current OTTV equations and coefficients were originally devel-
Life cycle cost technique had been used by Wong et al. [43] to oped for buildings with traditional bare roof. Up to now, there is no
analyze the economic benefit of green roofs. The study involved study conducted on this issue. It is a potential need to revise the
three roof types of a hypothetical building, namely the flat bare existing OTTV calculation method in order to involve this building
roof, extensive green roof and intensive green roof. Annual energy feature in the evaluation of OTTV.
consumption of the building was estimated by software Power- The objective of the present study is to revise the current OTTV
DOE. The study found that extensive green roof could achieve a calculation method, taking into account the reduced heat gain due
net saving of 14.6% which demonstrated a better option for build- to installation of green roof system in a building. A dataset of cor-
ing design under the climatic condition in Singapore. rection factors will be developed for computing the OTTVroof of a
Kosareo and Ries [44] extended the analysis to life cycle assess- building installed with green roof. The details of this study are re-
ment in the environmental impacts of the fabrication, transporta- ported in this paper.
tion, installation, operation, maintenance, and disposal of three
types of roof systems: conventional roof, intensive roof and exten- 2. Computer model and validation
sive green roof. The study revealed that, for the Pittsburgh PA cli-
mate, the energy use reduction was not very large in relation to 2.1. Experimental setup
the overall building energy use. Nonetheless, it was significant for
environmental impact over the life cycle of the building. Despite In 2009, a green roof system was constructed on the rooftop of a
the need for additional resources initially, green roof was the envi- commercial building in Hong Kong for the present study. As shown
ronmentally preferable choice when constructing a building. in Fig. 1a, three main types of vegetation with different growth
Peri et al. carried out life cycle costing and benefit cost analyses forms, height of plant and leaf area index (LAI) were planted in this
on green roof system, including the disposal cost of green roof sys- green roof space of area 45 m2 (7.5 m  6 m). The planted vegeta-
tem [45]. Their work allows a complete and proper application of tions include Zoysia Tenuifolia, Codiaeum Variegatum and Rhapis Ex-
the life cycle costing methodology in order to achieve an economic celsa. Table 1 lists the physical properties of these three types of
accounting of the disposal cost through the life cycle of a green roof vegetation used in this study. Below this green roof, there is an
system. air-conditioned office space with suspended ceiling installed.
In Hong Kong, Hui [46,47] had carried out investigations on The bare roof of this commercial building was constructed with
green roof technology with an aim to developing practical informa- four layers of traditional building materials including heavy con-
tion for its applications in Hong Kong. Study was carried out in a crete (inside layer), expanded polystyrene, cement screed and con-
number of building projects in which extensive green roof systems crete tile (outside layer). On top of this bare roof, the green roof
were installed. Temperatures of soil and ceiling void on the roofs was constructed. From bottom to top, this green roof system con-
were measured in summer days and compared to that of bare roof. sists of four layers of substrate including waterproof layer, drain-
Moreover, the characteristics and major design factors of different age layer, filter layer and lightweight soil. A manual irrigation
types of modular green roofs had been studied. Practical consider- system was installed. Daily irrigation was scheduled at 10:00 am
ations including durability of materials, estimation of system and 3:00 pm respectively.
weight and structural loading for soil substrates, supplemental irri- Arrays of temperature sensors were installed at different layers
gation, repair and maintenance for green roof systems had been of the substrate for measuring the hourly variation of temperature.
investigated. As shown in Fig. 1b and c, the sensors were fixed at the tile surface
Jim [48,49] had conducted research by experiment to develop a of the bare roof (i.e. at the bottom of the green roof setup); inside
cost-effective roof greening method for application in local build- the soil layer; and at the surface of the soil layer respectively. There
14 A.L.S. Chan, T.T. Chow / Applied Energy 107 (2013) 10–24

Signal
Cable

Temp.
Sensor

(a) (b)

Weather Station

Rhapis Excelsa
Codiaeum Variegatum
Zoysia Tenuifolia

Soil
Filter Layer
Drainage Layer
Waterproofing
Concrete Tile
Cement Screed
Expanded Polystyrene

Concrete Slab

: Temperature Sensor

: Heat Flux Sensor

(c)
Fig. 1. Green roof system installed on the rooftop of a commercial building (a) site photo of green roof (b) installation of temperature sensors and signal cables at construction
phase (c) schematic cross-sectional view of green roof.

Table 1 are 16 temperature sensors evenly installed at each layer of sur-

Physical properties of three vegetation types.
face. The distance between two adjacent sensors is about 1.5 m.
Plant species Zoysia Codiaeum Rhapis Moreover, a soil moisture sensor was buried into the soil layer (un-
Tenuifolia Variegatum Excelsa der the planted vegetation Zoysia Tenuifolia at a soil depth of
Height of plant (m) 0.035 0.6 1.1 10 cm) for monitoring the soil moisture content. Heat flux sensors
Leaf area index 2.2 3.5 5 were fixed at the undersurface of the concrete slab, directly below
Leaf reflectivity 0.29 0.26 0.2
the bare roof and green roof system. A pair of pyranometer was
Approximate percentage of occupied 60 20 20
area on the green roof space (%)
used to measure the reflectivity of leaf. The ranges of measurement
and accuracy of these sensors are listed in Table 2.
 A.L.S. Chan, T.T. Chow / Applied Energy 107 (2013) 10–24 15

A weather station was set up on the top of a pole behind the Summer Sunny Day Winter Sunny Day
green roof space to measure and record local meteorological data Summer Rainy Day Winter Rainy Day
including ambient air temperature, relative humidity, wind speed, 35
wind direction, rainfall and solar radiation. All sensors were con- 30

Air Temperature (o C)
nected to a data acquisition system by signal cables and the data
25
were recorded at an interval of 5 min.
20

2.2. Thermal performance of green roof system 15

10
The thermal performance of the bare roof and green roof system
5
in this study had been measured and monitored from July 2009 to
December 2010. It is recognized that the seasonal variation in 0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
weather condition has a significant impact on the thermal perfor- Hour
mance of a green roof system. Figs. 2–9 show the hourly profiles of
local weather parameters, surface temperature and heat flux Fig. 2. Hourly ambient air temperature measured on a sunny day and a rainy day in
through the bare roof and green roof measured on four sample summer season and winter season.
days in summer and winter seasons.

Summer Sunny Day Winter Sunny Day

2.2.1. Climatic condition in Hong Kong
Summer Rainy Day Winter Rainy Day
The climate in Hong Kong is subtropical, tending towards hot 120
and humid summers. As presented in Figs. 2 and 3, the air temper-
ature and relative humidity on a typical summer sunny day can 100

Relative Humidity (%)

reach up to 32 °C and 88% respectively. The solar radiation is rich 80
and the peak occurs at around 1:00 pm with a value of 995 W/
m2 (see Fig. 4). On a summer rainy day, the value of relative humid- 60
ity can rise to 99% while the air temperature is about 27 °C, as seen
40
from Figs. 2 and 3.
The winter season is relatively short and mild in this city. The 20
monthly mean air temperature is around 16 °C. Fig. 2 shows that
0
on a typical winter sunny day, the air temperatures range from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
10.7 °C to 18.3 °C. The fluctuation in relative humidity is large, Hour
ranging from 62% to 100% over a winter sunny day (see Fig. 3).
On a winter rainy day, the maximum air temperature is just about Fig. 3. Hourly relative humidity measured on a sunny day and a rainy day in
15 °C. The range of air temperature during a daytime period is be- summer season and winter season.

tween 12.4 °C and 14.2 °C. The values of relative humidity are high,
remaining 80% or above, over the whole winter rainy day.
Fig. 5, the values range from 24% to 26.8% on a winter rainy day;
and a summer rainy day gives a peak soil moisture content of
2.2.2. Soil moisture content 25.1% at 11:00 am and the minimum value of 23% occurs at
The hourly variation of soil moisture content in the green roof 12:00 midnight. The corresponding ranges of soil moisture content
system of this study was plotted in Fig. 5. It is observed that the soil for a winter sunny day and a summer sunny day are 22–23.3% and
moisture content is quite stable over a day, except some sudden in- 20.1–21.7%, respectively.
creases at around 10:00 am and 3:00 pm which are mainly caused
by scheduled irrigation. Moreover, the soil moisture contents on
winter days are slightly higher than that on summer days. It is be- 2.2.3. Heat flux
cause the solar radiation in the winter season is not as high as that The hourly variations of heat flux transmitted through a bare
in the summer season (see Fig. 4), leading to a relatively lower rate roof and a green roof on four sample days under summer and win-
of evapotranspiration which gives a higher soil moisture content ter seasons are plotted in Figs. 6 and 7 respectively. Positive values
remained. indicate heat gain into a building and vice versa. Since the vegeta-
It is not strange that rainy days give higher values of soil mois- tion plant Zoysia Tenuifolia occupies the largest portion (60%) of the
ture content than the sunny days due to rainfall. As depicted in green roof area in this study (as shown in Table 1), the measured

Table 2
Ranges of measurement and accuracy of sensors used in field measurement.

Sensor Range of measurement Accuracy

Thermal couple (RS type T) 200 to 250 °C ±0.0075  temp. measured
Heat flux plate (Hukse HFP01) +200 to 200 W/m2 ±5%
Soil moisture sensor (Delta-T ML2X) 0–0.5 m3/m3 ±0.01 m3/m3
Pyranometer (Kipp & Zonen CMP6) 0–1000 W/m2 ±1%
Weather station (Delta-T WS-HP1)
Air temperature 50 to 150 °C ±0.1 °C
Relative humidity 0–100% ±2%
Wind speed 0.2–75 m/s ±0.1 m/s
Wind direction 0–358 degrees ±0.3°
Solar radiation 0–2 kW/m2 ±15%
Rainfall Max: 500 mm in 1 h Sensitivity: 0.2 mm per tip
16 A.L.S. Chan, T.T. Chow / Applied Energy 107 (2013) 10–24

Summer Sunny Day Winter Sunny Day

Summer Rainy Day Winter Rainy Day
1200

Solar Radiation (W/m2)

1000

Global Horizontal
800

600

400

200

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour

Fig. 4. Hourly global horizontal solar radiation measured on a sunny day and a rainy day in summer season and winter season.

Summer Sunny Day Winter Sunny Day Winter Sunny Day - Bare Roof Winter Sunny Da y- Green Roof
Hour
Summer Rainy Day Winter Rainy Day Winter Rainy Day - Bare Roof Winter Rainy Day - Green Roof
30 0
Soil Moisture Content (%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
25 -2
Heat Flux (W/m2)
20
-4
15
-6
10
-8
5

0 -10
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour -12

Fig. 5. Hourly soil moisture content in a green roof measured on a sunny day and a Fig. 7. Hourly heat flux transmitted through a bare roof and a green roof measured
rainy day in summer season and winter season. on a sunny day and a rainy day in winter season.

Summer Sunny Day - Bare Roof Summer Sunny Day - Green Roof Summer Sunny Day - Bare Roof Summer Sunny Day - Green Roof
Summer Rainy Day - Bare Roof Summer Rainy Day - Green Roof Summer Rainy Day - Bare Roof Summer Rainy Day - Green Roof
12
Surface Temperature (oC)

60
10
50
8
Heat Flux (W/m2)

40
6
30
4
2 20

0 10

-2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
-4
Hour Hour

Fig. 6. Hourly heat flux transmitted through a bare roof and a green roof measured Fig. 8. Hourly surface temperature of a bare roof and a green roof measured on a
on a sunny day and a rainy day in summer season. sunny day and a rainy day in summer season.

data of this plant type were presented in this section for illustra- flux transmission ranges from 1.2 W/m2 (at 6:00 am) to 3.5 W/m2
tion and analysis. The measured results of all the three plant types (at 11:00 am).
(including Codiaeum Variegatum and Rhapis Excelsa) will be used On a summer rainy day, heat gain through both the bare roof
for model validation and detailed in Section 2.4 of this paper. and green roof are significantly reduced, showing its dependence
On a typical summer sunny day, the hourly variation of heat on solar radiation. The heat gain over daytime is fairly stable, giv-
flux gained into a building through a bare roof or a green roof fol- ing an average rate of heat gain of 1.9 W/m2 and 1.2 W/m2 for the
lows a similar trend of a typical cooling load profile of an air- bare roof and green roof respectively.
conditioned building, as shown in Fig. 6. In the bare roof, heat flux On a winter sunny day, heat loss from both the bare roof and
starts to rise from 8:00 am to a peak value of 9.7 W/m2 at 1:00 pm. green roof was observed. As presented in Fig. 7, there is heat loss
Then it declines gradually to nearly zero at around 1:00 am. With through the bare roof over the whole day, with a peak value of
the installation of a green roof, the heat flux into a building is sig- 6.6 W/m2 at 4:00 pm. A similar diurnal profile of heat loss was
nificantly reduced due to evapotranspiration of the vegetation and found for the green roof with comparatively higher magnitudes
additional insulation layers of the substrate. The reduction in heat than the bare roof. It is mainly caused by evapotranspiration of
 A.L.S. Chan, T.T. Chow / Applied Energy 107 (2013) 10–24 17

Winter Sunny Day - Bare Roof Winter Sunny Day - Green Roof tion through building envelope. Moreover, EnergyPlus can compute
Winter Rainy Day - Bare Roof Winter Rainy Day - Green Roof the building cooling/heating load and energy consumption of plant
equipment in an air-conditioning system, with varying indoor air
Surface Temperature (oC)

40
35 temperature of a building. A series of simulation tests had been
30 done on EnergyPlus according to the requirements of ANSI/ASHRAE
25 Standard 140-2011: Standard Method of Test for the Evaluation of
20
Building Energy Analysis Computer Programs [53–59].
15
The tests include simulations by EnergyPlus on the annual heat-
10
ing/cooling and peak heating/cooling for various building cases
5
with/without night ventilation, and with/without free floating
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 space temperatures. Moreover, the system/equipment perfor-
Hour mance of heating, ventilating & air-conditioning (HVAC) system
and the ability of EnergyPlus to predict zone loads, cooling coil
Fig. 9. Hourly surface temperature of a bare roof and a green roof measured on a
loads, cooling equipment energy consumption and resulting zone
sunny day and a rainy day in winter season.
environment were tested. The results of EnergyPlus were also com-
pared with the outputs from several other whole building energy
the planted vegetation in a green roof system. The heat loss profile analysis programs (such as ESP-r, BLAST, DOE2.1D, SRES/SUN,
shows that significant heat loss starts at 10:00 am, reaching a max- SERIRES, S3PAS, TRNSYS and TASE) that simulated the same test
imum of 7.3 W/m2 at 4:00 pm and then continues to decline to cases.
nearly zero at 8:00 am. It was found that the simulation results of EnergyPlus compared
Heat loss on a winter rainy day follows a similar diurnal pattern very closely with both analytical results and the simulation results
as that on a winter sunny day, but with increased magnitude in the from the eight whole building energy simulation programs. There-
absence of intensive solar radiation (see Fig. 4). Under a rainy day, fore EnergyPlus was adopted for building thermal and energy sim-
the rainfall causes a wet surface on the bare roof, resulting in a high- ulations in the present study.
er rate of convective heat loss than that through the green roof. In 2008, Sailor [41] had developed and integrated a model of en-
In the green roof area, the rate of heat loss under a rainy situa- ergy and moisture balance of vegetated rooftop into EnergyPlus.
tion is faster than that on a sunny day. It is because the solar radi- The model formulation is based on an Army Corps of Engineers’
ation is significantly reduced on a rainy day, leading to lesser heat FASST vegetation model [42] which solves simultaneously for the
gain transmitted into a building. As a result, heat loss through the foliage surface and soil temperatures at each time step. FASST
green roof is higher on a rainy day, even though the rate of evapo- tracks the energy and moisture balance within a vegetated soil. It
transpiration is reduced. is a one-dimensional model named Ecoroof that contains energy
budgets for both the foliage layer (Ff) and the ground surface (Fg).
The foliage energy balance is expressed as:
2.2.4. Surface temperature h i r e e r 
Figs. 8 and 9 present the hourly surface temperature profiles of f g f
F f ¼ rf I#s ð1  af Þ þ ef I#ir  ef rT 4f þ T 4g  T 4f þ Hf þ Lf
a bare roof and a green roof (Zoysia Tenuifolia) on a sunny day and a e1
rainy day in the summer season and winter season respectively. On ð1Þ
a summer sunny day, the surface temperature of a green roof var-
and the soil energy balance is given by:
ies with a range from 26.1 °C to 49.3 °C which is lower than that of
h i r e e r 
a bare roof over most of the hours (from 9:00 am to midnight), F g ¼ ð1  rf Þ I#s ð1  ag Þ þ eg I#ir  eg T 4g 
f g f
T 4g  T 4f
reflecting the thermal performance of a green roof system in the e1
presence of evapotranspiration and solar shading effect of the veg- @T g
þ H g þ Lg þ j  ð2Þ
etation (see Fig. 8). For the bare roof, maximum surface tempera- @z
ture of 56.5 °C occurs at 3:00 pm and the minimum temperature
This Ecoroof model includes moisture balance that allows pre-
24 °C is found at 5:00 am. The largest difference in surface temper-
cipitation, irrigation, and moisture transport between two soil lay-
ature (12.9 °C) between these two types of roof is found at
ers (see Fig. 10). As with a traditional bare roof, the energy balance
6:00 pm. From 1:00 am to 8:00 am, the surface temperature of
of a green roof is dominated by radiative heat transfer from the
the green roof is slightly higher than that of the bare roof.
sun. This solar radiation is balanced by sensible and latent heat flux
On a summer rainy day, the surface temperatures of both roof
from the soil and plant surfaces combined with conduction of heat
types drop drastically due to the absence of intensive solar radia-
into the soil substrate. From Eq. (1), in addition to convective and
tion and the convective cooling effect of rainfall. The temperature
sensible heat transfer, both the short and long wave radiation ab-
ranges of the bare roof and green roof are 17.3–28.4 °C and 18.1–
sorbed by the vegetation, including the effects of multiple reflec-
28.5 °C, respectively.
tions, have been taken into account. For energy budget at the soil
A similar trend of surface temperature profiles with reduced
surface, it is mainly influenced by the soil thermal properties, the
magnitude was found for both the bare roof and green roof under
amount of foliage coverage (rf) and the amount of moisture in
winter condition as illustrated in Fig. 9. On a winter rainy day, the
the soil. Eq. (2) represents the sensible heat flux (Hg), latent heat
surface temperatures of the bare roof are lower than that of the
flux (Lg) and the multiple reflections associated with long and short
green roof over the whole day, with a maximum difference of
wave radiation. The final term on the right side gives the conduc-
3.7 °C at 6:00 pm.
tion of heat into the soil substrate. With this Ecoroof model, Ener-
gyPlus can simulate and evaluate the thermal and energy
2.3. Computer simulation program and green roof model performance of a building constructed with a green roof system.

In the present study, a renowned building energy simulation 2.4. Validation of computer model
EnergyPlus was adopted. EnergyPlus is an energy analysis and
thermal load simulation program. It can compute and report hourly The Ecoroof model in EnergyPlus was validated by experimental
heat gain components including heat conduction and solar radia- results in the present study. For the purpose of model validation,
18 A.L.S. Chan, T.T. Chow / Applied Energy 107 (2013) 10–24

A positive value of MBE indicates that the simulation result is

higher than the experimental data and vice versa, while the RMSE
is a measure of how close the simulated value is to the experimen-
tal output. The calculated values of these two indices for the sur-
face temperature and heat flux of the bare roof and the green
roof systems (including Zoysia Tenuifolia, Codiaeum Variegatum
and Rhapis Excelsa) on four samples days (summer sunny day
(SSD), summer rainy day (SRD), winter sunny day (WSD) and win-
ter rainy day (WRD)) were listed in Table 3.
In terms of surface temperature, bare roof has comparatively
lower values of MBE, ranging from 1.4 °C to 2.1 °C while the mini-
mum and maximum of MBE for the green roof system are 1.6 °C
(Zoysia Tenuifolia on WSD) and 2.7 °C (Rhapis Excelsa on SSD),
respectively. The values of MBE (%) of surface temperature are less
than 10% for both the bare roof and green roof system, ranging
Fig. 10. Heat and moisture balance in a green roof system [60]. from 5.2% to 9.8%.
Another index, RMSE (%), is adopted to measure how close the
simulated value is to the experimental output. The results reveal
measured data from the green roof on four sample days including a that the bare roof has a better result (with a range of values from
sunny day and a rainy day in the summer and winter seasons were 6.3% to 10%) than the green roof (8.3–13.1%). Moreover, the green
used. During the testing, the hourly temperature of bare roof and roof systems Zoysia Tenuifolia and Codiaeum Variegatum have smal-
soil, the heat flux under the bare roof and green roof system, and ler values of RMSE (%) under sunny days (less than 10%) while that
the local meteorological data were measured and recorded. on rainy day reaches up to 12.1%. For the green roof with Rhapis
The office building installed with green roof was modeled using Excelsa, the values of RMSE (%) ranges from 10.9% to 13.1%.
EnergyPlus and its thermal performance was simulated with the For the comparison of heat flux, the minimum MBE (%) comes
use of measured local meteorological data. As there are three dif- from the case of bare roof on a winter sunny day (0.7%) and the
ferent types of vegetation planted into the green roof area of the case Zoysia Tenuifolia under a winter sunny day gives a maximum
office building, four building cases were modeled for the validation MBE (%) of 11.9%. In terms of RMSE (%), Codiaeum Variegatum per-
exercise including the bare roof, green roof (Zoysia Tenuifolia), forms better with a range of values from 9.4% to 11.3%. The values
green roof (Codiaeum Variegatum) and green roof (Rhapis Excelsa). of RMSE (%) of the bare roof, Zoysia Tenuifolia and Rhapis Excelsa fall
The physical properties of these three plant types such as height between 10.4% and 13.3%.
of plant, LAI and leaf reflectivity (as listed in Table 1) were used From this validation work, it is found that the computer simu-
as input data into EnergyPlus for building thermal and energy sim- lation program EnergyPlus with Ecoroof model perform adequately
ulations. The hourly profile of surface temperature and heat flux of and is reliable for predicting the thermal performance of a tradi-
the bare roof and the green roof (Zoysia Tenuifolia) measured on tional bare roof and a green roof system in the present study.
four sample days, as presented in Section 2.2, were plotted again
in Figs. 11 and 12, together with the corresponding simulation re- 3. Evaluation of OTTV for buildings with green roof systems
sults for illustration. Fig. 11 depicts the hourly profile of surface
temperature and Fig. 12 shows the hourly profile of heat flux. A The current OTTV equations and coefficients were originally
quantitative comparison between the measured and simulated re- developed for commercial buildings constructed with traditional
sults for all the four building cases including bare roof, green roof building façade. As indicated by Eqs. (5) and (6), there are two sets
(Zoysia Tenuifolia), green roof (Codiaeum Variegatum) and green of OTTV for a building: (i) OTTV for building external walls
roof (Rhapis Excelsa) are presented in Table 3. (OTTVwall) including the area of fenestration; and (ii) OTTV for a
As seen from Fig. 11, the simulation results over-estimated the building roof (OTTVroof) constructed with traditional building
surface temperature of both the bare roof and green roof in most materials such as tiles (outer layer), asphalt, cement, thermal
cases. Moreover, the hourly profiles of the simulation results are insulation, concrete and gypsum (inner layer). In a traditional bare
comparatively smooth while there are more fluctuations in the roof, the thermal transmittance (Ur) and absorptivity (ar) of the
measured surface temperatures, especially on the rainy days (see roof construction have a significant effect on the magnitude of
Fig. 11b, d, f and h). In comparison with the green roof, the bare OTTVroof.
roof gives smaller deviations between the simulated and measured
results. Nevertheless, the results from computer simulations follow ðAw  U w  aw  TDEQw Þ þ ðAf  SC  ESM  SFÞ
OTTV wall ¼ ð5Þ
the similar trend of the measured results. The hourly variations of Aw þ Af
heat flux presented in Fig. 12 give a similar analysis except that
there are slight underestimations of heat flux in the winter season ðAr  U r  ar  TDEQr Þ
OTTV roof ¼ ð6Þ
during early morning and late evening (see Fig. 12c, d, g and h). Ar
In order to compare the simulation results and measured data Nowadays, there is increasing application of green roof system
quantitatively, statistical indices namely mean bias error (MBE) on the rooftops of commercial buildings. One of its major functions
and root-mean-square error (RMSE) were used for error analysis is to reduce heat gain through the roof area as well as the cooling
of the simulation results. The definitions of MBE and RMSE are gi- requirement of a building in the summer season. A green roof typ-
ven by Eqs. (3) and (4) as shown below. ically consists of a substrate composite (water proof layer, drainage
Pn
 X exp;i Þ layer, filter layer and lightweight soil) and vegetation plant. Some
i¼1 ðX sim;i
MBE ¼ ð3Þ researchers had carried out study to evaluate the U-values of the
n
substrate layers. Niachou et al. [61] applied a steady state numer-
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Pn 2 ical approach to estimate the U-value of the substrate layers of a
i¼1 ðX sim;i  X exp;i Þ green roof system, while Kotsiris et al. [62] evaluate the U-value
RMSE ¼ ð4Þ
n of substrate components in a real scale under dynamic conditions.
 A.L.S. Chan, T.T. Chow / Applied Energy 107 (2013) 10–24 19

Fig. 11. Validation of Ecoroof model using simulated and measured surface temperatures of bare roof and green roof on a sunny day and a rainy day under summer and
winter seasons.

However, neither approach can determine an overall U-value of a 3.1. Correlation between OTTV and heat gain
green roof including the thermal insulation property of the vegeta-
tion plant. As a result, the original OTTV calculation method is not For evaluation of OTTV for buildings with green roofs, the first
appropriate for building constructed with green roof. A new ap- step is to develop a correlation between the OTTV of a traditional
proach should be developed for OTTV calculation of green roof, tak- bare roof construction (OTTVroof) and the sum of heat gain
ing into account the additional thermal insulation effect of a green through the roof area over the cooling season (from April to
roof system. October).
20 A.L.S. Chan, T.T. Chow / Applied Energy 107 (2013) 10–24

(a) Bare Roof on Summaer Sunny Day (b) Bare Roof on Summer Rainy Day

(c) Bare Roof on Winter Sunny Day (d) Bare Roof on Winter Rainy Day

(e) Green Roof on Summer Sunny Day (f) Green Roof on Summer Rainy Day

(g) Green Roof on Winter Sunny Day (h) Green Roof on Winter RainyDay
Fig. 12. Validation of Ecoroof model using simulated and measured heat fluxes of bare roof and green roof on a sunny day and a rainy day under summer and winter seasons.

Local architectural firms had been liaised for a survey on the signed and constructed in the recent ten years. From this survey,
design and construction of commercial buildings in Hong Kong. four types of typical and traditional bare roof construction were
Architectural drawings and design document were selected and identified. The details of construction materials and the values
collected for a total of 15 commercial buildings which were de- of OTTVroof of these four roof constructions were listed in Table 4.
 A.L.S. Chan, T.T. Chow / Applied Energy 107 (2013) 10–24 21

Table 3
Comparison of surface temperature and heat flux of bare roof and green roof between experimental data and simulation result.

Bare roof Green roof (Zoysia Tenuifolia) Green roof (Codiaeum Variegatum) Green roof (Rhapis Excelsa)
SSD SRD WSD WRD SSD SRD WSD WRD SSD SRD WSD WRD SSD SRD WSD WRD
(a) Surface temperature
MBE (°C) 2.0 2.1 1.4 1.6 2.3 2.4 1.6 2.5 2.5 2.3 1.8 2.4 2.7 1.9 1.9 2.2
MBE (%) 5.2 9.1 7.0 7.9 6.4 9.7 7.8 9.4 6.8 7.5 7.2 9.8 6.3 8.4 6.7 8.7
RMSE (°C) 2.4 2.2 1.5 2.0 2.9 2.3 1.5 2.7 2.4 2.6 1.9 2.7 3.0 2.7 2.1 2.4
RMSE (%) 6.3 10 7.6 8.0 8.3 10.1 8.7 12.1 9.4 11.2 9.6 12 11.6 10.9 11.2 13.1
(b) Heat flux
MBE (W/m2) 0.41 0.14 0.13 0.11 0.57 0.18 0.58 0.15 0.36 0.34 0.29 0.16 0.23 0.49 0.33 0.27
MBE (%) 9.2 9.8 0.7 2.6 9 11 11.9 0.9 8.2 8.6 10.3 6.8 7.5 7.8 11.1 9.3
RMSE (W/m2) 0.56 0.23 0.43 0.88 0.62 0.21 0.62 0.72 0.46 0.65 0.36 0.54 0.34 0.62 0.56 0.47
RMSE (%) 11 11.3 12.1 12.5 12.7 12.9 12.3 11.4 11.3 9.9 9.4 10.5 10.4 11.5 13.3 12.1

SSD: summer sunny day.

SRD: summer rainy day.
WSD: winter sunny day.
WRD: winter rainy day.

These four types of bare roof construction can be considered to The simulated results will be used to derive a dataset of appropri-
be the representative of majority of roof construction in air- ate correction factors for building designers to compute the
conditioned commercial buildings in Hong Kong. OTTVroof of a building installed with green roof system.
A generic fully air-conditioned office building was modeled Green roof system was incorporated into the four base case
using EnergyPlus. This office building is 40 stories high with a plan buildings established in this study and modeled by using Energy-
view of 36 m  36 m. It is an open plan office with four perimeter Plus. Four layers of basic components of a green roof system
zones facing North, East, South and West and a center zone. The including water proof, drainage, filter and soil layers were incorpo-
external wall was constructed with four layers of building rated on the top of the bare roof of the buildings as the arrange-
materials including mosaic tile, sand plastering, heavy concrete ment shown in Fig. 1c. Table 5 lists the physical properties of
and gypsum plastering. The floor-to-floor height is 3.4 m and a these four layers of substrate material.
window-to-wall ratio of 0.6 was set. In 2006, a consultant firm was commissioned by the local gov-
The four typical bare roof constructions identified above were ernment to conduct a study on green roof application in Hong Kong
incorporated into this generic office building so that four base [64]. In this study, a list of plant selection and soil thickness was
cases (A, B, C and D) were established and modeled by using Ener- derived from existing local knowledge and other overseas sources
gyPlus. For each base case, a set comprising 15 parametric simula- with humid climatic conditions similar to Hong Kong. These vege-
tion runs was performed with the use of a Hong Kong Typical tation types which have good chance for success were recom-
Meteorological Year (TMY) weather dataset [63]. In the parametric mended for planting in green roofs in Hong Kong. A vegetation
simulations, the values of tile absorptivity (0.1, 0.3, 0.5, 0.7 and 0.9) plant is mainly characterized by its value of LAI and height of plant.
and thickness of concrete slab (0.15, 0.2 and 0.25 m) of the roof By making reference to the recommendation of this consultancy
construction were varied sequentially. The sum of heat gain study, a series of parametric simulation runs had been performed
through the roof area over the cooling season (from April to Octo- for the four base case buildings with a combination of seven differ-
ber) for a given set of parameter values of the base case building ent soil thicknesses (from 0.1 m to 0.4 m), seven plant heights
was computed in each parametric run. A total of 60 multi-paramet- (from 0.05 m to 1.0 m) and eight different values of LAI (from 1.5
ric simulations had been carried out. to 5.0), giving a total of 1568 simulation runs. The values of pertur-
The results of the parametric simulations were illustrated by a bation for the parametric simulations are tabulated in Table 6.
plot of OTTVroof versus the sum of heat gain through the bare roof In each simulation run, a sum of heat gain through the green
construction as shown in Fig. 13. The regression curve does not roof area over the cooling season (from April to October) was com-
show a liner trend as commonly assumed. Since steady heat flux puted and represented as
is the product of U-value of a roof and the temperature difference
between indoor and outdoor space, it is expected that a linear rela- X
October
Q i;j;k;l;m :
tionship exists between the OTTVroof and the sum of heat gain. m¼April
However, taking into account that the direction of heat flow may
reverse at some hours during a 24-h diurnal cycle, heat loss where i is building case A, B, C or D, j is soil thickness (m), k is height
through the roof area may occur, especially for the roof construc- of plant (m), l is LAI, and m is calendar month (from April to
tion with higher value of OTTV. This results in a non-linear relation- October).
POctober
ship between the OTTVroof and the sum of heat gain. For instance, m¼April Q B;0:2;0:6;1:5;m is the sum of heat gain
Nevertheless, the regression results show a good correlation be- through a roof construction of building case B installed with green
tween the OTTVroof of the building case and the corresponding sum of roof system of soil thickness 0.2 m, plant height 0.6 m and LAI of
heat gain. The value of the correlation coefficient is 0.948. This cor- 1.5.
relation provides a means of determining correction factors for eval- The value of the computed heat gain was then input into the
uation of OTTVroof of commercial buildings constructed with green established correlation between OTTVroof and heat gain:
roof system. The details are presented in the following section. y = 15637 ln (x) + 24747 (developed in Section 3.1 and shown in
Fig. 13) to determine the corresponding value of OTTVroof for this
3.2. Development of correction factors for calculating OTTV of green building case installed with a green roof system, i.e. OTTVi,j,k,l. Then
roof a correction factor was derived by dividing the value of OTTVi,j,k,l
(for building case i with green roof) by OTTVi,BR (OTTV of roof for
The next step is to evaluate the thermal performance of green building case i with traditional bare roof only) as expressed in
roof constructions by employing EnergyPlus with Ecoroof model. Eq. (7) below:
22 A.L.S. Chan, T.T. Chow / Applied Energy 107 (2013) 10–24

Table 4
Construction details of four typical bare roof constructions identified through a survey.

Building case Material Thermal conductivity (W/mK) Density (kg/m3) Thickness (m) Absorptivity
A Concrete tile 1.1 2100 0.025 0.65
Asphalt 1.15 2350 0.02
Cement screed 0.72 1860 0.05
Expanded Polystyrene 0.034 25 0.155
Concrete 2.16 2400 0.15
Gypsum 0.38 1120 0.01 0.3
OTTVroof = 1.13 (W/m2)
B Concrete tiles 1.1 2100 0.025 0.65
Asphalt 1.15 2350 0.02
Sand screed 0.72 1860 0.05
Polystyrene 0.034 25 0.05
Concrete 2.16 2400 0.15
Gloss paint on gypsum 0.38 1120 0.01 0.3
OTTVroof = 2.50 (W/m2)
C Cement screed 0.72 1860 0.1 0.65
Expanded polystyrene 0.034 25 0.05
Concrete 2.16 2100 0.15 0.65
OTTVroof = 3.23 (W/m2)
D Granite marble slab 2.9 2650 0.02 0.65
Cement screed 0.72 1860 0.05
Expanded polystyrene 0.034 25 0.04
Concrete 2.16 2400 0.175 0.65
OTTVroof = 3.92 (W/m2)

50,000 Table 6
Values of perturbation for parametric simulations.
Sum of Heat Gain through

45,000
40,000
Building Soil thickness Height of plant Leaf area index
35,000 y = 15637ln(x) + 24747
Roof (kWh)

R² = 0.948 case (m) (m) (LAI)

30,000
25,000 A 0.10 0.05 1.5
20,000 B 0.15 0.1 2.0
15,000 C 0.20 0.2 2.5
10,000
D 0.25 0.4 3.0
0.30 0.6 3.5
5,000
0.35 0.8 4.0
-
0.40 1.0 4.5
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
5.0
OTTVroof (W/m2 )

Fig. 13. Correlation between OTTVroof and sum of heat gain.

lue of OTTVroof, the building designer can determine the category to
which the project building is belonged (i.e. category A, B, C or D).
Table 5 A green roof system is usually designed by a landscape archi-
Physical properties of substrate layers of a green roof system.
tect. The soil thickness can be recommended by an experienced
Thickness (m) Thermal conductivity (W/mK) gardener. Once the type of vegetation is selected, the correspond-
Soil layer 0.15 1.16 ing value of LAI and height of plant can be determined through lit-
Filter layer 0.001 0.3 erature survey or by field measurement.
Drainage layer 0.03 0.5 With these information/data of the project building and green
Waterproof layer 0.003 0.19
roof system, an appropriate correction factor can be selected from
the dataset developed in this study. Then the value of OTTVroof of a
building installed with green roof system can be determined by
OTTV i;j;k;l multiplying the OTTVroof of the bare roof by the selected correction
CF i;j;k;l ¼ ð7Þ
OTTV i;BR factor.

where CFi,j,k,l is a correction factor, OTTVi,j,k,l is the OTTV of a green

roof, and OTTVi,BR is the OTTV of a bare roof of a building case i. 4. Conclusion
Due to limited space in this paper, only one set of correction fac-
tors for building case D with soil thickness of 0.1 m is illustrated in Control on Overall Thermal Transfer Value (OTTV) of a building
Table 7. A full set of correction factors developed in this study are is a widely adopted measure in many countries and cities. There is
available through a hyper link as listed below (http://per- no doubt that regulating the OTTV of a building can enhance en-
sonal.cityu.edu.hk/~bsapplec/GR_OTTV_Correction_Factors.doc). ergy-efficient design of building envelope and mitigate the emis-
These predetermined correction factors can be used to compute sion of greenhouse gas. As the application of green roof system
the OTTV of a green roof of a commercial building. The derived into the rooftop of commercial buildings is becoming popular,
dataset of correction factors were divided into four categories there is a need to revise the existing OTTV calculation method that
according to the value of OTTV of bare roof (OTTVroof) of a building was originally developed for buildings with traditional bare roof
as indicated in Table 8. Building designers firstly calculate the va- constructions. In this study, a set of correction factors has been
lue of OTTVroof for the bare roof of a project building. With this va- developed for commercial buildings in subtropical Hong Kong.
 A.L.S. Chan, T.T. Chow / Applied Energy 107 (2013) 10–24 23

Table 7 [12] Guidelines for energy efficiency in buildings. Ministry of Energy,

Correction factors for building case/category D with soil thickness of 0.1 m. Telecommunications and Posts, Malaysia; December 1989.
[13] Janda KB, Busch JF. Worldwide status of energy standards for buildings. Energy
LAI Height of plant (m) 1994;19(1):27–44.
[14] Chirarattananon S, Taveekun J. An OTTV-based energy consumption model for
0.05 0.1 0.2 0.4 0.6 0.8 1
commercial buildings in Thailand. Energy Build 2004;36:680–9.
1.5 0.79 0.80 0.83 0.99 0.99 1.00 1.00 [15] Kunchornrat A, Namprakai P, Pont PT. The impacts of climate zones on the
2.0 0.50 0.50 0.52 0.65 0.78 0.93 0.96 energy performance of existing Thai buildings. Resour Conserv Recycl
2.5 0.30 0.30 0.31 0.41 0.54 0.65 0.85 2009;53:545–51.
3.0 0.20 0.21 0.21 0.28 0.37 0.48 0.62 [16] Chua KJ, Chou SK. Energy performance of residential buildings in Singapore.
3.5 0.15 0.15 0.15 0.19 0.25 0.38 0.49 Energy 2010;35:667–78.
[17] Chow WK, Chan KT. Overall thermal transfer values for building envelope in
4.0 0.10 0.11 0.11 0.12 0.19 0.31 0.40
Hong Kong. Appl Energy 1992;42:289–312.
4.5 0.07 0.07 0.09 0.09 0.15 0.22 0.34
[18] Hui SCM. Overall thermal transfer value (OTTV): how to improve its control in
5.0 0.05 0.05 0.06 0.06 0.08 0.18 0.30
Hong Kong. In: Proceedings of the One-day Symposium on Building, Energy
and Environment, Hong Kong; 16 October 1997. p. 12.1–12.11.
[19] Yik FWH, Wan KSY. An evaluation of the appropriateness of using overall
Table 8 thermal transfer value (OTTV) to regulate envelope energy performance of air-
Building category for correction factor selection. conditioned buildings. Energy 2005;30:41–71.
[20] Ho PHK. Modelling thermal performance of building envelope in Hong Kong. J
Building category Range of OTTVroof of bare roofa Build Surv 2000;2:21–6.
[21] Final report on the feasibility study on introduction of overall thermal transfer
A OTTVroof 6 1.82 W/m2
value (OTTV) to air conditioned buildings in Hong Kong. J. Roger Preston
B 1.82 W/m2 < OTTVroof 6 2.87 W/m2 Partners, vol. I and II; August 1991.
C 2.87 W/m2 < OTTVroof 6 3.58 W/m2 [22] Practice note for authorized persons, registered structural engineers and
D 3.58 W/m2 < OTTVroof registered geotechnical engineers APP-67. Energy Efficiency of Buildings,
a Building (Energy Efficiency) Regulation. Hong Kong SAR Government:
Values of OTTVroof forming the ranges are calculated by
Buildings Department; January 2011.
using the OTTVroof of the four building cases (refer to Table 4) [23] Wong NH, Chen Y, Ong CL, Sia A. Investigation of thermal benefits of rooftop
as follows: 1.82 = (1.13 + 2.5)/2; 2.87 = (2.5 + 3.23)/2; 3.58 = garden in the tropical environment. Build Environ 2003;38:261–70.
(3.23 + 3.92)/2. [24] Wong NH, Yok TP, Chen Y. Study of thermal performance of extensive rooftop
greenery systems in the tropical climate. Build Environ 2007;42:25–54.
These correction factors can be used by building designers to com- [25] Fang CF. Evaluating the thermal reduction effect of plant layers on rooftops.
Energy Build 2008;40:1048–52.
pute the OTTV of green roof constructed on the rooftop of a com-
[26] Ouldboukhitine SE, Belarbi R, Djedjig R. Characterization of green roof
mercial building, reflecting the thermal insulation capacity of components: measurements of thermal and hydrological properties. Build
substrate layers and vegetation plant in a green roof system. Environ 2012;56:78–85.
This study is being continued on the development of appropri- [27] Luikov AV. Heat and mass transfer in capillary porous bodies. Oxford:
Pergamon; 1966.
ate correction factors for buildings integrated with green walls. [28] Palomo E. Analysis of the green roofs cooling potential in buildings. Energy
Although the methodology of this study was adopted for subtrop- Build 1998;27:179–93.
ical Hong Kong, it is envisaged that this approach could be applied [29] Kumar R, Kaushik SC. Performance evaluation of green roof and shading for
thermal protection of buildings. Build Environ 2005;40:1505–11.
to other locations with different climatic conditions to derive [30] Theodosiou TG. Summer period analysis of the performance of a planted roof
appropriate correct factors for OTTV evaluation of green roof inte- as a passive cooling technique. Energy Build 2003;35:909–17.
grated buildings. [31] Alexandri E, Jones P. Developing a one-dimensional heat and mass transfer
algorithm for describing the effect of green roofs on the built environment:
comparison with experimental results. Build Environ 2007;42:2835–49.
Acknowledgement [32] Takakura T, Kitade S, Goto E. Cooling effect of greenery over a building. Energy
Build 2000;31:1–6.
[33] Onmura S, Matsumoto M, Hokoi S. Study on evaporative cooling effect of roof
The work described in this paper was fully supported by a grant lawn gardens. Energy Build 2001;33:653–66.
from the Research Grants Council of the Hong Kong Special Admin- [34] Takebayashi H, Moriyama M. Surface heat budget on green roof and high
reflection roof for mitigation of urban heat island. Build Environ
istrative Region, China [Project No. CityU 110809]. The authors are
2007;42:2971–9.
grateful to Ms. Esther Y T Mak and Mr. Harold H W Chan for their [35] Abalo SH, Banna M, Zeghmati B. Performance analysis of a planted roof as a
help in compiling the simulation results and statistical data. passive cooling technique in hot-humid tropics. Renewable Energy
2012;39:140–8.
[36] Niachou A, Papakonstantinou K, Santamouris M, Tsangrassoulis A,
References Mihalakakou G. Analysis of the green roof thermal properties and
investigation of its energy performance. Energy Build 2001;33:719–29.
[1] Code of Practice for Energy Efficiency of Lighting Installations. Hong Kong SAR [37] Lazzarin RM, Castellotti F, Busato F. Experimental measurements and
Government: Electrical and Mechanical Services Department; 2007. numerical modelling of a green roof. Energy Build 2005;37:1260–7.
[2] Code of Practice for Energy Efficiency of Air Conditioning Installations. Hong [38] Santamouris M, Pavlou C, Doukas P, Mihalakakou G, Synnefa A, Hatzibiros A,
Kong SAR Government: Electrical and Mechanical Services Department; 2007. et al. Investigating and analysing the energy and environmental performance
[3] Code of Practice for Energy Efficiency of Electrical Installations. Hong Kong SAR of an experimental green roof system installed in a nursery school building in
Government: Electrical and Mechanical Services Department; 2007. Athens, Greece. Energy 2007;32:1781–8.
[4] Code of Practice for Energy Efficiency of Lift & Escalator Installations. Hong [39] Spala A, Bagiorgas HS, Assimakopoulos MN, Kalavrouziotis J, Matthopoulos D,
Kong SAR Government: Electrical and Mechanical Services Department; 2007. Mihalakakou G. On the green roof system. Selection, state of the art and energy
[5] Code of Practice for Overall Thermal Transfer Value in Buildings. Hong Kong potential investigation of a system installed in an office building in Athens,
Government: Buildings Department; 1995. Greece. Renew Energy 2008;33:173–7.
[6] Performance-based Building Energy Code. Hong Kong SAR Government: [40] Wong NH, Cheong DKW, Yan H, Soh J, Ong CL, Sia A. The effects of rooftop
Electrical and Mechanical Services Department; 2007. garden on energy consumption of a commercial building in Singapore. Energy
[7] ASHRAE Standard 90-1975: Energy conservation in new building design. Build 2003;35:353–64.
Atlanta: American Society of Heating, Refrigerating and Air-conditioning [41] Sailor DJ. A green roof model for building energy simulation programs. Energy
Engineers; 1975. Build 2008;40:1466–78.
[8] Handbook on energy conservation in buildings and building services. Building [42] Frankenstein S, Koenig G. FASST Vegetation Models, U.S. Army engineer
and Construction Authority, Public Works Department, Singapore; 1979. research and development center, cold regions research and engineering
[9] Chou SK, Lee YK. A simplified overall thermal transfer value equation for laboratory (ERDC/CRREL). Technical Report TR-04-25; 2004.
building envelopes. Energy 1988;13(8):657–70. [43] Wong NH, Tay SF, Wong R, Ong CL, Sia A. Life cycle cost analysis of rooftop
[10] Guidelines for energy conserving design of buildings and utility systems. gardens in Singapore. Build Environ 2003;38:499–509.
Department of Energy, Republic of Philippines; 1993. [44] Kosareo L, Ries R. Comparative environmental life cycle assessment of green
[11] Chirarattananon S. Building energy efficiency in Thailand. In: Proceedings of roofs. Build Environ 2007;42:2606–13.
the symposium on energy efficient buildings. Hong Kong, 19 November 1992; [45] Peri G, Traverso M, Finkbeiner M, Rizzo G. The cost of green roofs disposal in a
p. 2.1–2.12. life cycle perspective: covering the gap. Energy 2012;48:406–14.
24 A.L.S. Chan, T.T. Chow / Applied Energy 107 (2013) 10–24

[46] Hui, SCM. Benefits and potential applications of green roof systems in Hong [55] Henninger Robert H, Witte Michael J. EnergyPlus testing with HVAC
Kong. In: Proceedings of the 2nd megacities international conference 2006, equipment performance tests CE100 to CE200 from ANSI/ASHRAE standard
Guangzhou, China; 1–2 December 2006. p. 351–60. 140-2011; October 2012.
[47] Hui, SCM. Development of modular green roofs for highly-density urban cities. [56] Henninger Robert H, Witte Michael J. EnergyPlus testing with HVAC
In: Proceedings of World Green Roof Congress 20008, London; 17–18 equipment performance tests CE300 to CE545 from ANSI/ASHRAE Standard
September 2008. p. 1–12. 140-2011; October 2012.
[48] Jim CY. Developing a cost-effective roof greening technology for energy [57] Henninger Robert H, Witte Michael J. EnergyPlus testing with fuel-fired
conservation. CLP Energy Innovation Fund, Hong Kong China Light & Power furnace tests HE100 to HE230 from ANSI/ASHRAE Standard 140-2011; October
(CLP) Holdings Limited. <https://www.clpgroup.com/SocNEnv/Env/EnvEdu/ 2012.
CLFund/Pages/2005SponsoredProjects.aspx>. [58] Henninger Robert H, Witte Michael J. EnergyPlus testing with HVAC
[49] Jim CY. Assessing a drought-tolerant plant (Stonecrops) for minimum- equipment component tests; October 2012.
maintenance green roofs in Hong Kong. CLP Energy Innovation Fund, Hong [59] Henninger Robert H, Witte Michael J. EnergyPlus testing with global energy
Kong China Light & Power (CLP) Holdings Limited. <https://www.clpgroup.com/ balance test; October 2012.
SocNEnv/Env/EnvEdu/CLFund/Pages/2006SponsoredProjects.aspx>. [60] EnergyPlus Development Team, Engineering Reference, EnergyPlus Version
[50] He H, Jim CY. Simulation of thermodynamic transmission in green roof 6.0; 2010.
ecosystem. Ecol Model 2010;221:2949–58. [61] Niachou A, Papakonstantinou K, Santamouris M, Tsangrassoulis A,
[51] Jim CY, He H. Coupling heat flux dynamics with meteorological conditions in Mihalakakou G. Analysis of the green roof thermal properties and
the green roof ecosystem. Ecol Eng 2010;36:1052–63. investigation of its energy performance. Energy Build 2001;33(7):719–29.
[52] Jim CY, Peng LLH. Weather effect on thermal and energy performance of an [62] Kotsiris G, Androutsopoulos A, Polychroni E, Nektarios PA. Dynamic U-value
extensive tropical green roof. Urban Forest Urban Greening 2012;11:73–85. estimation and energy simulation for green roofs. Energy Build
[53] Henninger Robert H, Witte Michael J. EnergyPlus testing with building thermal 2012;45:240–9.
envelope and fabric load tests from ANSI/ASHRAE Standard 140-2011; October [63] Chan ALS, Chow TT, Fong KF, Lin Z. Generation of a typical meteorological year
2012. for Hong Kong. Energy Convers Manage 2006;83(1):87–96.
[54] Henninger Robert H, Witte Michael J. EnergyPlus testing with HERS BESTEST [64] Study on green roof application in Hong Kong, Final Report. Urbis Limited; 16
tests from ANSI/ASHRAE standard 140-2011; October 2012. February 2007.