Roof is a unique building enclosure element due to its extreme environmental exposures and impact on whole building energy consumption. The roofing components affect the long-term environmental impact of the building. Embodied carbon of roofing component can be defined as the CO2 or GHG emission associated with extraction, manufacturing, transporting, installing, maintaining and disposing of roofing materials. Embodied carbon assessments of roofing (ECAR) draw on principles and methods from the field of life cycle assessment (LCA) to quantify the combined carbon emissions of each of these life cycle stages. The LCA assesses multiple environmental impacts across the full life cycle, whereas the ECAR focuses on a single impact category (e.g., global warming potential (GWP)) and applies a reduced scope (e.g., omitting the operational carbon). Furthermore, because of the roofing structure and assembly as well as the range of materials used whether in commercial or residential sectors, it is common to rely on collected data for the carbon emissions for each key component and processes using energy. This brief review states the current science related to the ECAR, points out the challenges and limitations of embodied carbon assessments and explains the data source of the ECAR research.
Embodied carbon assessment draws on techniques in the field of LCA. LCA of Roofing (LCAR) is a quantitative approach to describe and manage the environmental impacts of a roofing components across its entire life cycle through production, use, end-of-life treatment, recycling and final disposal. Thus, the LCAR indicates the multiple impact assessed from “cradle to grave” of roofing components. Meanwhile, the life cycle carbon assessments of roofing (LCCAR) determine the global warming potential (GWP) evaluated across full life cycle of roofing products. LCCAR focuses on GWP or carbon due in part to the emphasis on carbon emissions in environmental legislation. GWP is driven by emissions of greenhouse gases (GHGs) produced from human activities. The GWP of greenhouse gases is measured relative to CO2, using the unit of carbon dioxide equivalent or CO2e. For example, each kilogram of methane released into the atmosphere due to roofing production is estimated to have the equivalent GWP of 25 kilograms of carbon dioxide. Hence carbon factors which account for multiple greenhouse gases are reported in units of carbon dioxide equivalent, CO2e (IPCC, 2015).
The embodied carbon assessment of roofing (ECAR) also focuses on GWP but only on a single impact category across partial life-cycle such as one component production processes. Therefore, the difference between LCCAR and ECAR is whether to assess the impact of GWP across the full life cycle (LCCAR) or only on a partial life cycle of the processes (ECAR). Figure 1 shows how ECAR fits into the broader field of LCAR and LCCAR.
There are three main methods to conduct ECAR:
a) Process analysis (PA);
b) Environmentally extended input-output analysis (EEIOA);
PA is a bottom-up method, which has been developed to understand the carbon emission/embodied carbon of individual products. The method involves defining a product system with a boundary and creating an inventory of all the inputs and outputs between that product system and the environment. These flows can be materials or energy, and once the inventory has been established, the embodied carbon of each item in the inventory is evaluated. Here the focus is only on material level of roofing assemblies.
EEIOA is a top down approach, in that it uses macro-economic
data (national or regional input-output tables) as the basis for
assessing the environmental impact of a product. Input-output
tables ‘summarise economic transactions between sectors of an
economy’. By assigning environmental impacts such as carbon
emissions to the economic flows from each sector. Thus, it is
possible to estimate the total impacts of an output from any given
sector. EEIOA methods are most suitable for high level assessment
such as when considering emissions at a sector or national level,
thus it disregards the level of detail required to discriminate
between the impacts of individual products or components. The
reason for this is that economic input-output (IO) data is highly
The hybrid assessment combines the data from EEIOA models
with the data from process analysis (PA). However, recent research
has called into question the accuracy of hybrid assessment
methods, suggesting that the high levels of aggregation in EEIOA
models may lead to significant overestimation of the outcomes
(Yang et al., 2017).
Accordingly, of the three methods, process analysis (PA) is
to be most commonly applied to conduct the Embodied carbon
assessments of roofing. This may reflect the fact that this is the
method adopted in LCAR standards and the data and tools to
conduct process analysis of buildings are more widely available
than for EEIOA or HPA (Moncaster and Song, 2012).
a) Production processes are complex and subject to many
variations including sources of raw material, production volumes,
batching and etc. Thus data are based on averages.
b) Many production processes result in two or more coproducts
and emissions from the process should then be assigned
to each product in some meaningful way to avoid double counting.
c) Unless the assessment is conducted retrospectively,
at the end of the building life cycle, assumptions must be made
about future life cycle stages in order to account for all relevant
d) Data collection can be very resource intensive, meaning
simplifications may be necessary such as the use of secondary and
proxy data or estimates; the exclusion from the study of certain
parts of the building, or stages in the building life cycle; and the
use of cut-off criteria to exclude those materials or processes
deemed to contribute less than a specified threshold value.
e) The opportunity to reduce embodied carbon is greatest
in the early stages of design when the precise nature and quantum
of different material inputs are uncertain.
However, there are several data sources that should be
carefully investigated as follows.
An embodied carbon assessment of a roofing requires two
sets of data:
I. Inputs to and outputs from the roofing: For process
analysis, these take the form of physical quantities of materials
and energy at each life cycle stage. For EEIOA method, these data
are estimated based on economic transactions between sectors.
However, when the focus is only on the material level, these data
is better to be collected from manufacturers or tested in the lab.
II. The corresponding carbon factors for each of the
roofing processes: Carbon factors provide an estimate of the
total global warming potential (GWP) of all greenhouse gas
emissions associated with the production, use and disposal of
a given unit of material or the consumption of a unit of fuel or
electricity to produce roofing components for commercial or
For materials, carbon factors are typically reported in
either mass or volumetric terms with units of kgCO2e per kg
(or tonne) or per cubic meter (or other appropriate volumetric
unit) of material (Hammond & Jones, 2011). For energy, carbon
factors are reported either per mega. Joule or KWh of electricity,
or per physical unit of fuel consumed (e.g., litres) (Department
for Environment, Food and Rural Affairs, 2015). As for roofing
embodied carbon assessments, one can use the secondary sources
of carbon factor data since producing or obtaining primary data
would be too resource intensive (G. J. Treloar et al., 2000). A
variety of different sources of data exists including:
a) Literature derived carbon factors
These sources provide generic carbon factors for a range of
materials that have been derived by literature review. The methods
used to determine a single value from the range of values in the
literature vary between the different datasets. The secondary
data that they draw on include peer-reviewed, academic work
as well as data from industry and trade associations or specific
companies with regard to roofing components production, use,
assembly, and waste managing/cycling.
b) Industry data
Data that comes from industry bodies or individual companies
is released in the form of Environmental Product Declarations
(EPD) – e.g., carbon factors provided by a product supplier or
industry body. However, there are examples of non-EPD data from
industry for a range of materials. The World Steel Association
produces an extensive set of Life Cycle Inventory (LCI) data for
steel which is freely available for non-commercial use (World
Steel Association, 2011).
c) Government data
Government data are typically based on national inventories
or statistics meaning they are representative of typical carbon
emissions for materials or activities in a given country. Examples
of the use of Government data in embodied carbon assessments
of roofing included assessing emissions from electricity and fuel
consumption at different stages of the life cycle (Darby, 2014).
There are also examples of Government data sources for full
LCA inventories of materials, such as the US Life Cycle Inventory
Database (Azari-N & Kim, 2012). Moreover, when EEIO or hybrid
assessments are conducted, the input-output data are generally
derived from Government sources.
d) Factor from a commercial LCA database
Commercial LCA packages comprise a software interface for
conducting assessments and one or more databases of life cycle
inventory (LCI) and impact assessment data. LCI data are the basic
flows of materials and energy into and out of a product system.
The LCI assessment data are used to determine the potential
environmental impact of each of these flows.
Accordingly, the data quality needs to be assessed such
as the empirical reliability of the data, the robustness of data
acquisition methods, and the level of data validation. The use of
the data quality matrix involves scoring each data source against
the predetermined data quality criteria on roofing materials in
different sector whether commercial or residential.
In addition to carbon factor data, an embodied carbon
assessment requires data on the quantities of material used to
construct roofing components, and the fuels or electricity used in
the assembly and demolition. When an assessment is conducted
during the design stage of a roofing assembly (residential,
commercial, or vegetation roofing), materials quantities are
estimated from the design drawings and documents whilst energy
used in construction and demolition must be estimated.
For instance, first it is necessary to obtain material quantities
for the roofing assemblies. Then, fuel or electricity consumption
needs to be accounted for in the carbon factors for the relevant
life cycle stage for each material. Equation 1 is adapted from
Richardson et al. (2014) calculating embodied carbon and/or
energy for a roofing material. This equation represents embodied
carbon of initial roofing construction, embodied carbon of
subsequent refurbishment, and embodied carbon of demolition
respectively. The ECR refers to embodied carbon of the roofing
is the quantity of roofing material I;
is the carbon factor of material i per unit;
is the roofing construction energy requirement;
is the carbon factor of the roofing construction energy;
is the assumed service life of the roofing in years;
is the assumed service life of roofing components/
materials in years;
The process of extracting material quantities from design
documentation is known as a quantity takeoff (QTO). Traditionally
this has been conducted for the purposes of cost estimation
and was a manual process of visualisation and taking scaled
measurements from two dimensional technical drawings. The
development of BIM is seen as a further opportunity to improve
the efficiency of QTO processes (Cheung et al., 2012; Sattineni &
Potential advantages of adopting BIM processes have been
discussed and reviewed by academics, practitioners and policy
makers. Commonly cited benefits include capital and operational
cost savings, efficiency improvements, and increased profitability
in the design and delivery of roofing assemblies (McGraw Hill
Construction, 2014). Factors that are important for effective
implementation of BIM for embodied carbon assessment of
roofing components are ease of use and real-time appraisal.
The ease of use directly affect the amount of time required to
conduct an assessment. Real-time appraisal means that the effects
of different design changes on the embodied carbon of the roofing
assembly can be evaluated as the design is being developed. This
also ensures an efficient workflow, since the causal link between
design changes and negative or positive effects on the embodied
carbon results can be more readily appreciated. Thus, embodied
carbon impacts can be addressed before the design progresses
and changes become more difficult.
Creating a link between the BIM data and the carbon factor
data in order to prevent or minimise the amount of data reentry
and evaluating how suitable typical BIM models currently
are for this application are both key areas that require further
The use of a BIM integrated approach to conducting embodied
carbon assessments may improve their efficiency. Moreover,
if assessments can be conducted in near real-time during the
design, then BIM can facilitate more effective decision making
to reduce embodied carbon. However, the use of BIM in this way does not address some of the limitations of current methods.
These limitations cause uncertainty about the embodied carbon
results for the roofing. They also lead to uncertainty about the
reliability of comparative assessments of the embodied carbon
effects of different design options [1-18].
The following key conclusions drawn from the literature:
a) Process based embodied carbon assessment methods
provide valid and useful insights to quantify and reduce the
impact of the built environment on climate change;
b) BIM is seen by some as a useful source of material
quantity data for conducting embodied carbon assessments of
c) Tools have been developed to exploit the data generation
capabilities of BIM for the purpose embodied carbon assessments.
But this approach is relatively new and how suitable BIM data are
for this purpose and how best to link BIM data and carbon factor
data requires further exploration;
d) However, an apparent lack of comparability of results
between studies is viewed as problematic. This is generally
acknowledged to be due to variations in methods and data used;
e) The availability of carbon factor data at the product level
has been highlighted as a cause of variation;
f) Despite the perceived lack of comparability and the
acknowledgement of uncertainty about the outcomes of embodied
carbon assessments, formal uncertainty assessment was only
applied in only several of studies;
g) The studies of embodied carbon of buildings where
uncertainty assessments were conducted lack a comprehensive
review of the relevant sources of uncertainty;
The methods applied are predominantly quantitative and
where the scope of the uncertainty assessment has been restricted
to selected parameters or scenarios, the justification for their
selection is unclear.