Studying construction materials flows and stock: A review

Full text

Authors’version (post-print) of the article :
Augiseau, V., & Barles, S. (2017). Studying construction materials flows and stock: A review.
Resources, Conservation and Recycling, 123, 153-164.
Studying Construction Materials Flows and Stock: A Review
Author names and affiliations
Vincent Augiseau, Ph.D. Candidate
Paris 1 University
UMR CNRS 8504 - Géographie-Cités - CRIA
13 rue du Four, 75006 Paris, FRANCE
augiseau@parisgeo.cnrs.fr
Sabine Barles, Professor
Paris 1 University
UMR CNRS 8504 - Géographie-Cités - CRIA
13 rue du Four, 75006 Paris, FRANCE
sabine.barles@univ-paris1.fr
Corresponding author
Vincent Augiseau
Highlights
− Review of thirty-one publications dedicated to construction materials
− Methodological survey of studies analysing jointly flows and stock
− Focus on the major construction materials flows: non-metallic minerals
− Identification of six main methodological approaches
− Synthesis of the main quantitative results in various contexts
Summary
Thirty-one scientific publications on the joint study of construction materials flows and stock with a
focus on non-metallic minerals are reviewed. These studies serve different purposes: forecasting and
comparing future input and output flows, studying the influence of several parameters on future flows,
estimating the present or future stock as well as its evolution, studying urban metabolism and
analysing the interaction between flows and stock. They are carried out at national, regional or urban
level and their time scale range from a century to a single year.
Six main methodological approaches can be distinguished: static bottom-up or top-down flow
analysis; bottom-up stock analysis; dynamic retrospective or prospective flow analysis using flow-
driven or stock-driven models; and top-down prospective or retrospective stock analysis using a flow-
driven model. Approaches are often combined, which is a way to accounting for uncertainty. They rely
on assumptions such as homogeneity of material composition and lifetime within groups of built
works, whereas quality and coverage of data used are very variable.
Most of the case studied show that stock accumulation is still ongoing and that non-metallic mineral
secondary resources would be insufficient to totally meet future demand. They also point out
infrastructures as the major part of the stock. Reviewed studies contributed to the development of a

methodological framework for the joint study of flows and stock, as well as a conceptual framework
for analysing the metabolism of a socioeconomic system. Further research could develop these
frameworks and support the implementation of industrial ecology policies.
Keywords
Construction Materials; Non-Metallic Minerals; MFA; Stock and Flow Accounting Methodology
1. Introduction
Construction materials are the largest flows entering urban areas after water, while they constitute the
top waste deposit (Adriaanse et al., 1997; Matthews et al., 2000). The consumption of these mostly
non-renewable materials strongly increased since the mid-20th century in most urban areas (Douglas
and Lawson, 2002; Kennedy et al., 2007). It generates many environmental impacts, from extraction,
transformation, transportation, to end-life management and especially storage. Moreover, the
expansion of a city strongly constrains local mineral resources extraction. This leads to the extension
of the areas supplying the city as reported for Toronto by Kennedy et al. (2007), and hence to
increasing costs and environmental imprint. It also leads to emerging scarcity at a local scale, as in the
case of Geneva (Rochat et al., 2006). Output flows to landfills also raise environmental issues together
with land-use conflicts.
Materials contained in a city today, in the form of buildings and networks, could potentially be
recycled tomorrow, through urban mining and so partly substitute for primary resources in highly
urbanised countries (Brunner, 2011). Moreover anthropogenic stock is a physical link between the
demand for resources (inputs) and waste (outputs). Hence, it should not be considered as being only
determined by flows but also as a driver of these flows (Müller, 2006). The joint analysis of flows and
stock is therefore an important issue in terms of understanding and managing the metabolism of
socioeconomic systems as defined by Fisher-Kowalski (2011). This is bound up with significant
methodological challenges concerning the knowledge of flows and stock of existing materials, in
terms of quantity and quality, along with short-term forecasting which is essential to anticipating and
acting on metabolism.
Research conducted since the late 1990s had different purposes and focused on a variety of spatial and
time scales, built works and material. It led to the development of methods dedicated to construction
materials flows and stock accounting which differ in terms of assumptions and data used. The aim of
this paper, based on the review of literature, is to analyse these methods. After having introduced the
selected studies (Section 2), their purposes and scope are analysed and compared (Section 3) assuming
that both of them can influence the methodological choices, so the latter can’t be studied without those
contextual elements. Section 4 deals with the methods themselves: main approaches that can be
identified (Section 4.1), assumptions (4.2.) and data (4.3.) which are used, and how uncertainty is
taken into account (4.4.). This analysis is followed by a summary of key results (Section 5) and leads
to a discussion (Section 6) on the issues that the methodological framework (6.1.) as well as the
conceptual framework (6.2.) used in the studies raise and how further research could address them.
The field of industrial ecology actions that flows and stock analysis could support, in addition to
recycling, is then discussed in Section 6.3.
2. Selected studies
Studies were selected according to a two-step method. A review of scientific publications on flows and
stocks of construction materials, or which estimate such flows and stock as part of urban metabolism
studies, was first carried out. It was followed by a restricted selection according to four criteria:
- Joint studies of construction materials flows and stock: works focusing only on flows or on stock,
even if interesting, were not retained. Due to their numbers, studies giving only estimates of current
waste flows generated by construction activity were not included, in particular those using generation
rate calculation methods such as developed by Yost and Halstead (1996). Thirty-six studies using this
method have already been reviewed by Wu et al. (2014).

- Studies relating to non-metallic minerals or several building materials (non-metallic minerals and
metals or wood or other materials): a comprehensive review of sixty dynamic analysis studies of
metals flows and stock was published by Müller et al. (2014). In addition, thirteen of the twenty-five
material stock studies reviewed by Tanikawa et al. (2015) relate exclusively to one or more metals. In
order to complement these two articles, studies addressing other materials and so-called static analysis
were selected in this paper.
- Publications in English: preference was given to selecting English-language articles. However
several studies written in French were included to make this research better known.
- Recent publications: articles published since 2000 have been favoured to present the current state of
research.
Thirty-one publications were thus selected. A systematic analysis was then carried out according to
seven main questions:
− What is the purpose of the study?
− Which spatial and time scales, built works, materials and dissipative flows are considered?
− Which methodological approach is applied for studying current flows and stock?
− Which methodological approach is applied for forecasting?
− What are the main assumptions used by such approach?
− Which key data are used?
− How is uncertainty taken into account?
Table 1 presents the selected studies according to the first six questions: purpose, scope, main
approach for studying current flows and stock, main approach for forecasting, assumptions, and data.
Publications are sorted by year and authors’ names. Studies carried out by researchers from the same
University or Institute are grouped (and links indicated).
Table 1. Summary presentation of selected studies
Studies
Purpose
Space and time
scales, built works,
materials and
dissipative flows
Main methodological
approach for
studying current
flows and stock
Main
methodological
approach for
forecasting
Assumptions and data
Obernosterer
et al. (1998)
Studying urban
metabolism
Estimating the
present stock
Urban (Vienna in
Austria)
1991
All constructions
All materials,
metals in detail
(iron, aluminium,
zinc, lead)
Bottom-up stock
analysis
Static bottom-up flow
analysis
No forecasting
Kolher and
Hassler
(2002),
Kohler and
Yang (2007),
Yang and
Kohler
(2008)
Forecasting and
comparing future
input and output
flows
Studying the
influence of several
parameters on
future flows
Estimating the
future stock
National (Germany
and China)
1992, until 2030
(Germany); 2005,
since 1978, until
2050 (China)
Buildings; road
(including bridges
and tunnels), rail,
water, gas and
electricity networks
Cement, steel,
wood, bricks, sand,
aggregates,
bitumen, glass,
plastics, non-
ferrous metals
Static bottom-up flow
analysis
Static top-down flow
analysis
Dynamic flow
analysis: stock-
driven model
Assumptions: stock depends
on population size,
urbanisation rates and
national infrastructure
development; future
construction, renovation and
demolition according to
national objectives or
correlated to tendency over
the last 20 years
Data: material intensity as
well as energy and
associated environmental
impacts for 12 types of
buildings and 6 periods of
construction based on a life-
cycle analysis of 50 building
elements (bottom-up);
imports, exports and
domestic extraction of
minerals (top-down)
Faist
Studying urban
Urban (canton of
Bottom-up stock
No forecasting
Assumptions: no more

Studies
Purpose
Space and time
scales, built works,
materials and
dissipative flows
Main methodological
approach for
studying current
flows and stock
Main
methodological
approach for
forecasting
Assumptions and data
Emmenegger
,
Frischknecht
(2003)
metabolism
Estimating the
present stock
Geneva in
Switzerland)
2000
Buildings; road,
rail, water, gas and
electricity networks
Gravel/sand,
concrete, bitumen,
bricks, plastics,
wood, iron, copper,
aluminium
analysis
Static bottom-up flow
analysis
networks development
Data : statistics on
construction, renovation and
demolition according to 10
types; material content for 6
types of buildings and for
networks; demolition rates
for each type of building;
renewal rate or lifetime by
type of network
Müller et al.
(2004),
Müller
(2006)
Forecasting and
comparing future
input and output
flows
Studying the
influence of several
parameters on
future flows
Regional
(Kreuzung
Schweizer
Mittelland in
Switzerland) and
national
(Netherlands)
1900-1997, 1997-
2100 (Kreuzung
Schweizer
Mittelland); 1900-
2003, 2003-2100
(Netherlands)
Housing
Wood (Kreuzung
Schweizer
Mittelland);
Concrete and wood
(Netherlands)
Retrospective dynamic
flow analysis: stock-
driven model
Prospective dynamic
flow analysis: stock-
driven model
Assumptions: housing stock
depends on population size
and lifestyle (useful floor
area per capita); constant
services provided by housing
units during their lifetime
Data: national statistics on
population and (until
present) housing stock;
average useful space in
housing units and number of
persons per unit (until
present); concrete intensity
in buildings and national
concrete production
Bergsdal et
al. (2007),
Sartori et al.
(2008),
Brattebø et
al. (2009),
Sandberg et
al. (2014a,
2014b)
(Norwegian
University of
Science &
Technology)
Forecasting and
comparing future
input and output
flows
Studying the
influence of several
parameters on
future flows
National (Norway)
1900-2005, 2005-
2100
1800-2011, 2011-
2050 (Sandberg et
al., 2014a, 2014b)
Housing; bridges
(Brattebø et al.,
2009)
Concrete and wood
+ PCB (buildings);
steel and bitumen
(bridges) (Brattebø
et al., 2009)
Same as Müller (2006)
Same as for current
flows and stock
No forecasting for
bridges (Brattebø et
al., 2009)
Assumptions: same as
Müller (2006)
Data : same as Müller
(2006); + survival functions
for housing renovation flows
based on lifetime before
renovation (Sartori et al.,
2008; Sandberg et al.,
2014a, 2014b); composition
in concrete and wood for 5
or 2 types of housing and 9
or 5 types of age classes
(national statistics, studies
and exchanges with
professionals)
Hashimoto et
al. (2007,
2009)
Forecasting and
comparing future
input and output
flows
National (Japan)
1995, since 1970,
until 2030
Buildings; road,
rail, water, gas
electricity and
telecommunications
networks; public
works (ports,
airports, dikes)
Non-metallic
minerals
Dissipative flows
study
Bottom-up stock
analysis
Static bottom-up flow
analysis
and static top-down
analysis
Retrospective dynamic
flow analysis: flow-
driven model
Prospective dynamic
flow analysis: flow-
driven model,
survival probability
Assumptions: future annual
construction equals the
average of the last 10 years
Data: annual construction of
buildings, networks, roads
and public works since 1970,
material intensities based on
public statistics and data
from trade associations
(bottom-up); domestic
extraction (top-down)
Schiller
(2007),
Deilmann
(2009)
Estimating the
present stock
Forecasting and
comparing future
Urban and national
(3 towns in Saxony
in Germany, and
Germany)
Bottom-up approach
for stock
Prospective dynamic
flow analysis: flow-
driven model
Assumptions: strong
influence of urban fabrics on
material demand; future
housing construction
according to demographic

Studies
Purpose
Space and time
scales, built works,
materials and
dissipative flows
Main methodological
approach for
studying current
flows and stock
Main
methodological
approach for
forecasting
Assumptions and data
(Leibniz
Institute of
Ecological
Urban and
Regional
Development
)
input and output
flows
1995-2001, until
2025 (Saxony);
2000, until 2025
(Germany)
Housing; road and
water networks
Concrete, stone,
bitumen, steel, iron
and plastics
forecasts
Data: material intensities
associated to different types
of urban fabric and building
techniques, rate of estate
renewal, rate of material
recycling
Kapur et al.
(2008)
Estimating the
present stock
Studying the stock
evolution
National (United
States)
1900-2000
Buildings; road and
water networks
Cement
Retrospective dynamic
flow analysis: flow-
driven model
No forecasting
Data: national consumption
of cement, distribution of
consumption according to 8
uses
Lichtensteige
r, Baccini
(2008)
Estimating the
present stock
National
(Switzerland)
2000, since 1900
Buildings
Gravel-sand, marl-
clay, cement, wood
and copper
Dissipative flows:
loose gravel-sand
and copper
Top-down
retrospective stock
analysis using a flow-
driven model
No forecasting
Assumptions: total
construction during the 20th
century equals the stock in
2000 (building renewal rate
of 0.1 %)
Data: annual construction
since 1900; production of
cement since 1902 and
imports and exports of
copper
Barles (2009,
2014)
Studying urban
metabolism
Regional and urban
(Paris and Paris
region, Midi-
Pyrénées, Ariège,
Haute-Garonne in
France)
2003 and 2006
All constructions
All materials; non-
metallic minerals in
detail
Dissipative flows
related to road
surface wear
Static top-down flow
analysis
No forecasting
Data : freight statistics, land
use and construction
statistics
Daxbeck et
al. (2009)
Estimating the
present stock
Forecasting future
output flows
National (countries
of the European
Union)
2003, until 2035
Buildings; road,
rail, water, gas,
electricity, heating
and
telecommunication
networks
Concrete/stone,
non-ferrous metals,
glass, plastics, iron,
wood, bricks, slag
Bottom-up stock
analysis
Prospective dynamic
flow analysis: flow-
driven model
Assumptions: future
construction activity
correlated with national
economic activity (GDP)
Data: surfaces, material
intensities and lifetimes per
period of construction (6)
Hu et al.
(2010a,
2010b)
Forecasting and
comparing future
input and output
flows
Studying the
influence of several
parameters on
future flows
National and urban
(China, Beijing)
1900-2006, 2006-
2100 (China);
1949-2005, 2005-
2050 (Beijing)
Housing
Same as Müller (2006)
Same as for current
flows and stock
Assumptions: same as
Müller (2006); + stock of
housing depends on
population size, lifestyle and
wealth (useful floor area per
capita according to
GDP/inhabitant); urban and
rural stocks as two sub-
systems linked by migration

Studies
Purpose
Space and time
scales, built works,
materials and
dissipative flows
Main methodological
approach for
studying current
flows and stock
Main
methodological
approach for
forecasting
Assumptions and data
Concrete, iron and
steel
flows
Rouvreau et
al. (2012),
Serrand et al.
(2013)
(BRGM,
French
Geological
Survey
Office)
Estimating and
locating flows
Estimating the
present stock
Forecasting and
comparing future
input and output
flows (Serrand et
al., 2013)
Urban (Orléans in
France)
2004-2006, until
2030
Buildings; road,
rail, water, gas and
electricity networks
Non-metallic
minerals, metals,
wood, glass,
plastics, bitumen
Bottom-up stock
analysis
Static bottom-up flow
analysis
Prospective dynamic
flow analysis: flow-
driven model
Assumptions: homogenous
areas within a city in terms
of buildings material
intensity; future housing
construction according to
local development plan
Data: building permits,
geographical databases and
aerial photos
Shi et al.
(2012),
Huang et al.
(2013)
(author or co-
author of
both articles)
Forecasting and
comparing future
input and output
flows
Studying the
influence of several
parameters on
future flows
National (China)
1950-2010, 2011-
2050
Buildings; road and
rail networks (Shi
et al., 2012)
Cement,
aggregates, sand,
lime, bricks,
bitumen, steel,
glass, wood
Same as Hu et al.
(2010a, 2010b)
Same as for current
flows and stock
Assumptions: same as Hu et
al. (2010a, 2010b); + road
and rail networks
development depends on
wealth (GDP per capita) and
population density
Fishman et
al. (2014)
Studying the stock
evolution
(saturation
phenomena)
National (United
States and Japan)
1930-2005, 2005-
2050
All constructions
Non-metallic
minerals, wood,
iron, other metals
Top-down
retrospective stock
analysis: flow-driven
model
Top-down
prospective stock
analysis: flow-
driven model
Assumptions: stock volume
as the result of annual net
addition to stock multiplied
by stocking rates; future
stock addition equals the
average addition over the
last 10 years or correlated to
demographic forecasts
Data: imports, used domestic
extraction and exports since
1870 and 1878; stocking
rates of materials
(estimations by authors)
Tanikawa et
al. (2015),
Fishman et
al. (2015)
(author or co-
author of
both articles)
Estimating the
present stock
Studying the stock
evolution
National and
regional (Japan and
47 prefectures in
Japan)
2010, 1945-2010
(Japan), 1965-2010
(prefectures)
Buildings; road, rail
and water
networks, public
works (ports,
airports, dikes)
Aggregates,
cement, iron,
asphalt, wood,
other
Bottom-up stock
analysis
Top-down
retrospective stock
analysis: flow-driven
model
No forecasting
Assumptions: stock
evolutions dependent on
population, GDP per capita
and material intensity
(stock/GDP) (Fishman et al.,
2015)
Data: geographical
databases, annual
construction, legal
construction codes for
material intensities
Wiedenhofer
et al. (2015)
Estimating the
present stock
Forecasting and
comparing future
input and output
flows
National (25
countries of the
European Union)
2004-2009, until
2020
Buildings; road and
rail networks
Non-metallic
minerals, asphalt,
Bottom-up stock
analysis
Static bottom-up flow
analysis
and static top-down
flow analysis
Prospective dynamic
flow analysis: flow-
driven model
Assumptions: future annual
renovation, construction and
demolition equals the
average of the last 7 or 10
years, increasing recycling
rate
Data: material intensities
according to 72 types of
buildings depending on
climate zones

Studies
Purpose
Space and time
scales, built works,
materials and
dissipative flows
Main methodological
approach for
studying current
flows and stock
Main
methodological
approach for
forecasting
Assumptions and data
bitumen
3. Purposes and scopes of the studies
Though they are expressed in various ways, four main purposes can be identified. Forecasting and
comparing future input and output flows is the purpose of twenty-one of the publications which are
seeking to anticipate the consumption of materials (especially of aggregates) or the generation of
waste. These studies aim to define future recycling potential using a comparison of inflows and
outflows. This purpose can be reached through a methodological approach that enables studying the
influence of several parameters on future flows, which is the case for fifteen of those studies.
Estimating the present stock is the aim of fifteen studies seeking to identify the characteristics of the
anthropogenic stock in terms of material composition. Some of them also focus on studying the stock
evolution: they adopt a retrospective approach and use data about the age and location of the stock
(Kapur et al., 2008; Fishman et al., 2014; Fishman et al., 2015; Tanikawa et al., 2015). Fishman et al.
(2014) also adopt a prospective approach so as to analyse a potential stock saturation phenomenon in
countries that are already highly urbanised. Estimating the future stock composition is the purpose of
five studies seeking to forecast the evolution of the anthropogenic stock, in particular in countries
experiencing major urban development. Studying urban metabolism, including flows of construction
materials, is the aim of four studies. They aim at understanding the metabolism of urban areas,
including the interrelation between a city and its hinterland, and to establish a diagnosis prior to
industrial ecology policies. Data on the characteristics of stock are used to estimate, analyse and
compare certain flows as for Barles (2009, 2014) and the composition of the stock is also studied by
Obernosterer et al. (1998) and Faist Emmenegger and Frischknecht (2003). A fifth subject, which is
cross-cutting to the above is the analysis of the interaction between flows and stock, especially of
factors determining stock accumulation or removal.
Spatial scales are very mixed, ranging from cities with tens of thousands inhabitants in Saxony
(Deilmann, 2009), to all the countries of the European Union (Daxbeck et al., 2009; Wiedenhofer et
al., 2015), or large countries like China (Yang and Kohler, 2008; Hu et al., 2010a; Shi et al,. 2012;
Huang et al., 2013) and the USA (Kapur et al., 2008); Fishman et al., 2014). Twenty-four studies are
conducted at the national level, three of them dealing with more than one country (Daxbeck et al.,
2009; Fishman et al., 2014; Wiedenhofer et al., 2015). Five studies focus on a regional scale (Müller
et al, 2004; Barles, 2009, 2014; Tanikawa et al., 2015; Fishman et al., 2015) and eight on an urban
scale: cities or wider urban areas (Obernosterer et al., 1998; Faist Emmenegger, and Frischknecht,
2003; Deilmann, 2009; Barles, 2009, 2014; Hu et al., 2010a; Rouvreau et al., 2012; Serrand et al.,
2013).
Time scales are also very variable, ranging from retrospective and prospective studies covering a
century, to an assessment for a single reference year. A prospective horizon for a century is used for
analysing the likely trends in the evolution of flows, including phasing between inflows and outflows,
often at the national level as for Müller (2006). A closer time horizon of fifty years is also considered
so as to limit uncertainties (Sandberg et al., 2014a and 2014b). By contrast, local studies (Deilmann,
2009; Serrand et al., 2013) adopt a horizon of about twenty years, which is more compatible with
regional and urban planning time frameworks.
The scope of built works studied is much linked to spatial and temporal scales: the broader these are
the more selection is restricted. Thus most national forecasting studies focus on one or two types of
built works and relate to one or two materials. Ten studies focus on buildings and with the exception
of Lichtensteiger and Baccini (2008) on housing in particular (Müller et al., 2004; Müller, 2006;
Bergsdal et al., 2007; Sartori et al., 2008; Hu et al., 2010a, 2010b; Huang et al., 2013; Sandberg et al.,
2014a, 2014b). Housing is often studied because it is better covered by public statistics, unlike non
domestic buildings. Fifteen studies take into account networks, five of them considering only two
kind: road and rail for Shi et al., (2012) and Wiedenhofer et al. (2015); road and water networks for

Schiller (2007), Kapur et al. (2008), and Deilmann (2009). Bridges are studied by the publications
directed by Kohler and also by Brattebø et al., 2009. Studies about the Japanese stock also consider
civil works like ports, airports and dikes (Hashimoto et al., 2007, 2009; Tanikawa et al., 2015;
Fishman et al., 2015). Urban metabolism studies cover all or most of built works and materials.
Materials which are most studied are those with most mass, "bulk materials", in the terminology of
Lichtensteiger and Baccini (2008, p. 42). These include especially non-metallic minerals: aggregates
(or gravel), sand and concrete. Other rarer substances called "trace materials" in this same terminology
include non-ferrous metals and copper in particular. Wood, iron and steel are also the subject of
several comprehensive studies, unlike other materials such as stone, bricks and tiles, plaster, glass,
plastics, bitumen. The study of concrete in a retrospective approach can restrict the time scale to the
late 19th and early 20th centuries, when its use became widely distributed in conventional constructions
(Müller, 2006 and further studies on Norway and China; Kapur et al., 2008). It also allows estimation
of the consumption of concrete components (aggregates, sand, cement and metal reinforcements),
assuming mixtures and reinforcement that are identical at a national level.
Dissipative flows as defined by Ayres (1994) are not as often the subject of estimation in the case of
non-metallic minerals for construction, as in the studies of metallic materials reviewed by Müller et al.
(2014). However, Hashimoto et al. (2009) propose a typology and an estimation of dissipative flows
by crossing different methods and data sources on the output flows. Authors distinguish between:
"potential wastes and secondary resources" that have a high probability of becoming wastes or of
being recovered; "potentially dissipated materials" that are unlikely to emerge as wastes or secondary
resources due to their dissipation during or after use; "dissipatively used materials", such as aggregates
for levelling a construction site; and "permanent structures" that have a low probability of being
demolished (Hashimoto et al., 2009 p. 2860). Dissipative losses of gravel, sand and copper are taken
into account by Baccini et al. (2008), whereas losses related to road surface wear are estimated in
Barles (2009), by assuming an average annual wear per road meter.
4. Methods applied in the studies
4.1. A combination of six main approaches
Four methodological approaches are usually distinguished for material flow analysis: bottom-up or
top-down on the one hand, static or dynamic on the other. Similarly, two approaches are commonly
defined for the analysis of the stock: bottom-up or top-down (Birat et al., 2013.). In studies relating to
flows and stock, these approaches are combined, while the detail of methods varies highly, according
to data used and assumptions.
The bottom-up approach to material flow analysis, as defined by Brunner and Rechberger (2004) is
based on a prior definition of processes in which materials circulate, and then proceed in estimating
flows, process by process. This approach allows all flows to be traced, from the extraction of minerals
to their transformation, use and recovery. It therefore allows detailed identification of actions to be
taken (Brunner and Rechberger, 2004).
Definitions of the top-down approach to material flow analysis for a socioeconomic system have been
given by Eurostat (2001), and relating to urban areas by Barles (2009). It involves the study of flows
without considering a priori the internal processes of the system. The domestic material consumption
in an area and the net addition to stock are generally calculated. The joint study of flows and stock is
not the primary purpose of this approach, in which stock is considered as a "black box", and the net
addition to stock mechanism is not always sufficiently analysed (Matthews et al., 2000). That said,
comparing the apparent consumption of materials to the annual construction of housing and to the
extension of the urban area, allows the distinction between buildings and networks in the net addition
to stock (Barles, 2014).
A material flow analysis can also be performed with a static approach, that is to say within a reference
period which is usually a year, or using a dynamic approach as defined by Baccini and Bader (1996).
In the latter case, the change in flows over a long period is studied by assuming removal from the
stock of materials contained in built works which reached the end of their lifetime. An average lifetime
or a mathematical survival function (probability for a built work to be demolished after a given

number of years) is used. Dynamic analysis can be based on input flows, for example by extrapolating
their recent yearly average: such models are usually called flow-driven and also "demand-driven
modelling" (Tanikawa et al., 2015, p. 779). Alternatively, stock-driven models are based on the
assumption that “the stock of “service units” is the driver for the material flows. The stock can be
estimated by an assigned “development pattern” […] or “stock expansion rate” […], or it can be
defined as a function of the population and its lifestyle” (Hu et al., 2010b, p. 442).
The study of material stock may also be done using two approaches that can be described as bottom-
up or top-down. The bottom-up approach is based on a division of the stock into categories (housing,
business premises, etc.), and then by the application of material ratios or intensities (in tonnes/m² for
example). It can also be described as static. Elements in the stock are generally differentiated
according to a typology that crosses criteria relating to function, form and age: for example housing
units in a building of less than three floors built between 1945 and 1960. This approach provides a
good knowledge of the "inner structure" of the stock, and allows both the quantity and the quality of
materials to be taken into account (Lichtensteiger and Baccini, 2008, p. 46).
The top-down approach is to quantify stock as the sum of annual net additions to stock over a long
period. Stocks are thus derived from the difference between inflows and outflows, calculated from
year-to-year. These flows are known from statistical data (construction and demolition), or are
estimated, based on average lifetimes or survival functions. This top-down approach to stock analysis,
or "top-down accounting" according to Tanikawa et al. (2015, p. 779) is analogous to the dynamic
flow analysis using a stock-driven model approach defined above, differing only in its purpose.
Based on the distinction between flow analysis and stock analysis on the one hand, and the
assumptions used in modelling on the other hand, three main approaches for static analysis and three
for dynamic analysis can be identified:
- Static bottom-up flow analysis: eight studies (Obernosterer et al., 1998; Kolher and Hassler, 2002;
Faist Emmenegger and Frischknecht, 2003; Hashimoto et al., 2007, 2009; Kohler and Yang; 2007;
Yang and Kohler, 2008; Wiedenhofer et al., 2015);
- Static top-down flow analysis: ten studies (Kolher and Hassler, 2002; Faist Emmenegger and
Frischknecht 2003; Hashimoto et al., 2007, 2009; Kohler and Yang; 2007; Yang and Kohler, 2008;
Barles, 2009, 2014; Rouvreau et al., 2012; Wiedenhofer et al., 2015);
- Bottom-up stock analysis: twelve studies (Obernosterer et al., 1998; Faist Emmenegger and
Frischknecht, 2003; Hashimoto et al., 2007, 2009; Kohler and Yang; 2007; Schiller, 2007; Yang and
Kohler, 2008; Deilmann, 2009; Rouvreau et al., 2012; Fishman et al., 2015; Tanikawa et al., 2015;
Wiedenhofer et al., 2015);
- Dynamic retrospective or prospective flow analysis using a flow-driven model (input flows) : eight
studies (Hashimoto et al., 2007, 2009; Schiller, 2007; Kapur et al., 2008; Daxbeck et al., 2009;
Deilmann, 2009; Serrand et al., 2013; Wiedenhofer et al., 2015)
- Dynamic retrospective or prospective flow analysis using a stock-driven model: twelve articles
(Kolher and Hassler, 2002; Müller et al., 2004; Müller, 2006; Bergsdal et al., 2007; Kohler and Yang;
2007; Sartori et al., 2008; Yang and Kohler, 2008; Hu et al., 2010a, 2010b; Huang et al., 2013;
Sandberg et al., 2014a, 2014b);
- Top-down retrospective or prospective stock analysis using a flow-driven model: four publications
(Lichtensteiger and Baccini, 2008; Fishman et al., 2014; Fishman et al., 2015; Tanikawa et al., 2015).
This typology corresponds to that established by Müller et al. (2014), in their review of dynamic flow
analysis methods for metals: “Retrospective top-down”; “Retrospective and prospective top-down;
“Prospective top-down” (applied to metals in emerging technologies and based on assumptions such
as demand growth rates); “Retrospective bottom-up” and “Retrospective and prospective bottom-up”.
Top-down and bottom-up approaches for flow analysis are often associated, and Kohler and Yang
(2007) recommend combining them. The authors consider that top-down approach provides an upper
limit to flows and bottom-up approach a lower limit, while making it possible to locate and date the
flows. A static bottom-up analysis may also use a dynamic approach to estimate certain flows resulting

from the maintenance of networks, especially roads through renewal rates and average lifetimes (Faist
Emmenegger and Frischknecht, 2003).
In addition, the results of bottom-up stock analyses are often used to carry out a prospective dynamic
flow analysis with a flow-driven model. The outflows resulting from the demolition of the part of the
stock regarded as being at the end of its lifetime are then taken into account. Furthermore, top-down
prospective stock analysis using a stock-driven model often proceeds with a bottom-up analysis of the
stock on a past period. This analysis carried out in a first step is used to calibrate the model. For
example, Müller (2006) estimates the material intensity of concrete from 1900 to 2003 by measuring
the concrete content of typical representatives of buildings categories.
4.2. Assumptions
An important assumption which is shared by static bottom-up flow analysis, bottom-up stock analysis,
as well as dynamic flow analysis, is that built works can be divided into groups or types which have
the same material intensity. Criteria such as construction period, use of a building (e.g. residential or
non residential) or location (e.g. neighbourhood for Rouvreau et al., 2012 or climate zone for
Wiedenhofer et al., 2015) are set so as to distinguish built works. Schiller (2007) and Deilmann (2009)
also consider that urban fabrics have a strong influence on material intensity and they analyse
buildings and networks at the level of “urban structural types”. Dynamic flow analysis methods also
assume that each group of built works has an average lifetime and they express the probability of a
building to be demolished with survival functions. Müller et al. (2014) make a detailed review of these
mathematical functions.
Dynamic prospective flow analysis and top-down prospective stock analysis using a flow-driven
model generally assume that future construction, and sometimes also renovation and demolition, are
correlated to factors such as national or local planning objectives (Yang and Kohler, 2008; Serrand et
al., 2013), national or local demographic forecasts (Fishman et al., 2014; Schiller, 2007; Deilmann,
2009), or economic activity and especially GDP forecasts (Daxbeck et al., 2009). The tendency over
the last years can also be used and sometimes combined with the previous factors, the period taken
into account varying from seven or ten years (Hashimoto et al., 2007, 2009; Fishman et al., 2014;
Wiedenhofer et al., 2015) to twenty years (Kolher and Hassler, 2002; Kohler and Yang; 2007; Yang
and Kohler, 2008).
Dynamic retrospective or prospective flow analysis using a stock-driven model relies on assumptions
about the relation between flows and stock. According to Müller (2006), neither prices nor GDP
determine the stock. Instead, population and its lifestyle generate a demand for services which are
provided by the stock. When applied to the case study of housing, the service unit is the useful floor
area, and the lifestyle variable is the useful floor area per capita. Another important assumption is that
constant services are provided by housing units during their lifetime. Studies by Kohler use population
size and specific surface/network per person rates as variables. The model originally developed by
Müller also considers population size and useful floor area per capita and sets two exogenous variables
to estimate housing demand: household size (number of persons per family, so by approximation per
housing unit) and average floor area of housing units. These variables are used to deduct the useful
floor area per capita. Sandberg et al. (2014a) use the number of housing units as a parameter, because
they consider data on the average floor area per housing units as highly uncertain. Therefore, demand
for housing is assumed to depend on population and housing occupancy rates (number of persons per
dwelling). Further studies on China also take into account the influence of GDP per capita on the
housing floor area per capita (Hu et al., 2010a, 2010b) and on road and rail networks development
(Shi et al., 2012; Huang et al., 2013).
Top-down retrospective or prospective stock analysis uses very specific assumptions: stock volume is
defined as the result of annual net additions to stock multiplied by average stocking rates (Fishman et
al., 2014), whereas Lichtensteiger and Baccini (2008) consider the Swiss building stock in 2000 as
equal to the total construction during the 20th century (building renewal rate of 0.1 %).

4.3. Data
Bottom-up methods use data from administrations controlling building permits which provide
information on the declared floor areas of buildings constructed in a year. Their reliability and
accuracy vary by jurisdiction and recorded dates are often those when building permission requests are
filed, whereas works begin later (or do not even start), and may last for longer than one year. Taking
into account the two years preceding the year studied, and then calculating an average over the period
solves this problem (Barles, 2009). Demolition and renovation of buildings are often not subject to
monitoring by the public authorities as systematically as is construction. This is also true to the
construction, renovation and demolition of networks. Moreover, existing data on these activities are
often not subject to linear or volume quantification, but are expressed in monetary terms (micro
economic data).
Top-down flow analysis is usually based on statistics on the production and transport of materials
either produced by administrations (including macro economic data such as input-output tables) or by
organisations representing materials producers. They are available at the national or regional level and
are often organised by product categories, where the use of such goods for construction or other
activities is not always distinguished.
Data on existing buildings floor areas and networks lengths are often included in geographic
information systems produced by state agencies or private companies in the developed countries. Four
sources of uncertainty can however be identified according to Rouvreau et al. (2012): building
surfaces are often the outer surfaces of roofs and not those of the walls; when heights are indicated,
they often indicate average heights: surfaces constructed in basements or underground cannot be
identified, as well as the use of some buildings (e.g. housing and offices not distinguished). Similarly
different types of roads may not be distinguished (Tanikawa et al., 2015).
Statistics on the date or period of construction of buildings are also necessary for dynamic approaches
but may be inaccurate, especially for buildings constructed before the mid-20th century. If data are
recorded by time periods, the latter may be too broad and so not allow some changes in building
techniques to be considered. To validate or supplement statistics, an estimate using historical aerial
photographs and field visits is useful though it can usually be performed only on a limited spatial
perimeter (Rouvreau et al., 2012; see also Tanikawa and Hashimoto, 2009). When statistics are
missing or unreliable, using satellite images to produce spatial data presents an alternative (Hsu et al.,
2013; Liang et al., 2014).
Material intensities come from two sources: case studies or expert opinion on the amount or
distribution of materials in demolition waste on the one hand; modelling buildings and networks on
the basis of architectural studies, construction standards, or expert opinion on the other hand. Built
works lifetimes are also based on case studies or expert opinion as in the case of Hashimoto et al.
(2007). Building elements lifetimes are defined by Sandberg et al. (2014a) to estimate flows generated
by renovation. Information on stock can also be provided by tax or public health data which was not
the case for these reviewed studies.
4.4. Uncertainty
Some studies do not take uncertainty into account, at least explicitly. This is the case for most of the
first studies on urban metabolism. When the question of uncertainty is addressed, three approaches can
be distinguished.
Using intervals or rounding results is especially used for bottom-up approaches, in particular in stock
analyses. For example, Daxbeck et al. (2009) indicate data used in estimating the material composition
of buildings and the resulting estimates, with averages and intervals. Tanikawa et al. (2015) round the
results of their stock estimates, and present them at a higher spatial scale greater than one kilometre
square. Other studies estimate variations in results as a percentage. This is the case, for example, of the
bottom-up study of stock by Rouvreau et al. (2012), which estimate the average uncertainties
associated with surface areas and material intensities.

Crossing or comparing results from different methodological approaches is a second way to deal with
uncertainty. Indeed, several studies cross bottom-up and top-down approaches. According to Yang and
Kohler (2008), uncertainty problems "can only be overcome through a systematic modelling of
missing data by combining different and often contradictory data with plausible assumptions, and by
validating the general model by comparing methodologically independent top-down and bottom-up
results" (Yang and Kohler, 2008, p. 6).
Finally, sensitivity analysis is used in retrospective or prospective dynamic flow analysis using a
stock-driven model so as to test the impact of changes in input parameters on the results, according to
different scenarios. A sensitivity analysis is also carried out by Fishman et al. (2014).
This analysis of ways to accounting for uncertainty is consistent with the results of Müller et al. (2014)
who identify four approaches: i) data uncertainty is not considered; ii) sensitivity analysis, with
different average lifetimes or different lifetime distributions and standard deviations; iii) uncertainty
intervals; and iv) Gaussian error propagation to calculate the standard deviation of stock and flows,
based on standard deviations that are defined for each input variable and parameter. This last approach
has not been observed here.
It can be noticed that uncertainty is being increasingly taken into consideration. For example,
Tanikawa et al. (2015) in their bottom-up analysis of the stock, identify and analyse the influence of
three sources of uncertainty: data collection and typology, material intensity factors, match of each
type to its material intensity. They also compare their results to those obtained from another bottom-up
approach (Hashimoto et al., 2007) and from a top-down stock analysis (Fishman et al., 2014).
5. Some key results of the studies
Studies show that flows of construction materials through European urban areas remain significant not
only in terms of consumption but also in terms of processed output. Indeed, whereas annual domestic
material consumption ranges from 2.6 t/cap in Paris region (Barles, 2014; minerals only), 3.9 t/cap in
Geneva (Faist and Emmenegger, 2003), and up to 6.5 t/cap in Vienna (Obernosterer et al., 1998), the
total input to output ratio (including excavated materials) reaches 1.5 in Paris region, 3.5 in Geneva
and 1.4 in Vienna. Annual net addition to stock remains also large: 1.1 t/cap in Paris region, to 2.8
t/cap in Geneva and 5.5 t/cap in Vienna. Non-metallic minerals accounts for more than 90 % of those
flows in weight for urban areas like Orléans, Geneva and Vienna.
The in-use stock on non-metallic minerals reaches 244 t/cap in Orléans (Rouvreau et al., 2012), 294
t/cap in Japan, 337 t/cap in the USA (Fishman et al., 2014), and 375 t/cap in Switzerland (Baccini et
al., 2008) (built works included varies). In-use cement stock alone reaches 14.7 t/cap in the USA
(Kapur et al., 2008). In Japan, in 2010, 43 % of the materials are stocked in buildings, 26 % in roads,
19 % in seaports and 8 % in dams (Tanikawa et al., 2015). In the European Union (EU25), non-
metallic minerals stock in roads reaches 128 t/cap, 72 t/cap for residential buildings, and 3 t/cap for
railways in 2009 (Wiedenhofer et al., 2015).
Besides, the distinction between buildings and networks in the consumption of materials helps to
better connect metabolism and urban development patterns. Indeed, material stock in networks is
about 20 % of total stock in heavily built-up areas while it reaches two-thirds of the in-use stock in
low-density areas (Deilmann, 2009). The study conducted on Germany by Schiller (2007) shows that
the annual consumption of materials could be reduced by 25 % within twenty-five years (2000-2025),
thanks to a higher urban density (compared to a trend scenario). In contrast, a larger share of
consumption would be dedicated to the maintenance of networks in the case of low-density urban
development. Using indicators such as the domestic material consumption of building materials by
floor area built in the year on the one hand, and extension of the urban area in a year per each new
inhabitant on the other hand shows the specific dynamics of material consumption within an urban
area (Barles, 2014). Indeed, these two indicators are higher in the outskirts of Paris region which
experience urban sprawl than they are in the centre area where networks are already developed.
Hashimoto et al. (2009) estimate dissipative flows in Japan and show that: i) the mass of potential
waste and secondary resources is equal to 30 % of the material used for the construction of the works;
ii) potentially dissipated materials account for 0-5 % respectively; iii) materials used in a dissipative

manner represent 45 % to 50 % of the total; and iv) that permanent structures account for 15 % to 20
% of all materials.
Prospective dynamic flows analysis indicates that even if recycling rates increase, secondary resources
could only marginally substitute for primary resources in cities like Vienna or Orléans. In Vienna,
improved recycling of construction waste would only reduce the use of primary resources by 7 %
(Obernosterer et al., 1998). In Orléans, over a period of twenty-five years (2005-2030), secondary
resources of minerals would only cover 27 % of the aggregates needed in this area, even if 70 % of
output materials are recycled (Serrand et al., 2013). In the UE25, increased recycling would only cover
half of the materials needed for ongoing expansion of the stock of residential buildings, roads and rail
networks (Wiedenhofer et al., 2015).
Studies using a stock-driven model show the influence of stock on the future needs for primary
resources, and allow anticipation of a delay between future construction and demolition activities
(Müller, 2006; Bergsdal et al., 2007; Sartori et al., 2008; Sandberg et al., 2014a, 2014b). Indeed, in
countries with a mature urban development as the Netherlands and Norway, renovation activity would
be more important than construction in the first decades of the 21st century. Major demolition would
then take place in the second half of the century, which would generate a need for construction to
maintain the stock of built works. Construction and demolition are therefore set to be “phase
displaced” (Müller, 2006, p. 154) at the beginning of the 21st century, but would subsequently evolve
together. In most scenarios, non-metallic mineral secondary resources would thus be insufficient to
meet demand. Between 34 % to 58 % of the domestic material consumption of non-metallic minerals
in the EU25 in 2009 are used for maintenance and this share is expected to increase (Wiedenhofer et
al., 2015). Thus, Yang and Kohler (2008) consider that in Europe the existing stock built since 1850
will largely determine flows in the next forty years.
In China, building stock increased by a factor of 5.5 between 1978 and 2005, it should double in the
next thirty years and then stay constant (Yang and Kohler, 2008). Present construction techniques
could lead to a rise in the output flows from 2030 due to short lifetime of buildings. This would also
lead to a peak in material inputs that would occur around 2030 or 2035 (Yang and Kohler, 2008;
Huang et al., 2013). As a consequence, secondary resources will not supply future demand for
construction materials in cities like Beijing (Hu et al., 2010b).
Some studies also provide a better understanding of factors determining stock accumulation.
Sensitivity analysis carried out by Sandberg et al., (2014b) shows the impact on flows of two main
parameters: population size and dwellings’ lifetime Fishman et al. (2015) use an IPAT model where
environmental impact (I) is considered as the consequence of three drivers: population (P), affluence
(A), and technology (T). Applied to the case of Japan since 1945, this model points out the various
influences in time and space of these factors on stock accumulation. Studies also provide some
indications of the presence or absence of stock saturation phenomena. Fishman et al. (2014)
demonstrate that stock accumulation is still ongoing in 2005 in developed countries like Japan and the
United States. But saturation is probable in the future with “dematerialisation of total stock” in Japan
and “dematerialisation of per capita stock” in the USA (Fishman et al., 2014, p. 418).
6. Discussion
6.1. Improving the methodological framework
Research conducted since the late 1990s led to the development of a wide variety of methods for
studying construction materials flows and stock. These methods serve different purposes, whether
focusing on flows or stock, estimating them for a single year or on a longer time scale (retrospective or
prospective). But they mostly depend on data: according to data availability, different space and time
scales, built works, and flows can be taken into account and different levels of precision can be
reached. The quality and coverage of data are very variable from one country to another, and even
from a city or region to another. Hence it seems impossible to define a “standard” method for flows
and stock analysis. But some principles may be formulated in terms of definitions (inflows, outflows,
stock), indicators and methodological approach, as for instance crossing bottom-up and top-down
approaches to compare their respective results.

Then, the review shows that most studies focus on buildings and do not take sufficient account of
infrastructures, i.e. networks and civil works like ports and dams. Tanikawa et al. (2015) and
Wiedenhofer et al. (2015) show that these constitute the major part of the stock in Japan and the
European Union. Hence they raise important issues in terms of inflows to maintain them today and in
the coming years. Research should better address these issues when data are available. The fact that
infrastructures are partly located underground increases data collection difficulty as it makes
impossible the use of methods like remote sensing.
Besides, the assumptions and data used by bottom-up approaches still raise questions. Firstly, the
assumption of homogeneity within different types of built works is problematic when buildings or
networks grouped within the same type according to their use can vary a lot. This is especially true for
non-residential buildings as for instance offices, logistics centres, shopping centres or sports centres.
Recent research on non domestic buildings stock in Germany addresses this issue (Ortlepp et al.,
2015). In addition, the assumption of homogeneity for works built during the same period leads to
rather approximate estimates, notably for periods before the mid-20th century, and it does not usually
incorporate the changes caused by renovation. Furthermore, homogeneity is generally assumed across
national urban areas, while building techniques can vary significantly between regions given local
resources, climate, history, and other factors. Studying the material composition at a neighbourhood
level is an interesting approach to link the composition of buildings to the one of networks, but it also
raises the problem of homogeneity of urban structural types.
Besides, material intensities rely on case studies, expert opinion or modelling and significant
differences can be observed according to sources. This is especially true for trace materials (e.g.
copper), but also for non-metallic mineral waste in the inclusion or not of excavated materials
(Kleemann et al., 2014). Average lifetimes are also problematic. Indeed, while construction periods
are generally considered crucial to the quality of the built works, and hence for the choice of
refurbishment or demolition, according to Kohler and Hassler (2002, p. 232): "there is no relation
between the age or condition of an individual building and the probability that it will be demolished".
As a consequence, local case studies should be pursued, in the first place to produce more accurate
estimates of material intensities and lifetimes. Relevant crossing of different data sources and of top-
down and bottom-up approaches can also enhance the reliability of estimates. Producing data from
actual practices of renovation and demolition that can be observed is preferable. Methods for studying
the stock survival by the analysis of local statistics on demolition, such as that defined by Bradley and
Kohler (2007) or the recent analysis of building mortality patterns and reasons for demolition by
Aksözen et al. (2016) provide a relevant solution. Building morphology analysis as defined by
Steadman (1994) or the combination of remote sensing and cadastre data (Grün et al., 2009; Quinn,
2012) that enables to relate large number of georeferenced information with historical data could also
bring interesting results in the future. Furthermore, modern construction history as gained importance
since the end of the 20th century
i
and could provide additional information about the stock.
Besides, material flow analysis approaches establish mass flows for country, urban area or groups of
buildings whereas life-cycle assessment (LCA) and life-cycle costing (LCC) use functional units
relating to required functions, area or volume, occupancy, and reference period. LCA and LCC use a
hierarchical decomposition of buildings by elements which allows encapsulating material data from
life-cycle inventories. Similar basic data are used by MFA, LCA and LCC, so that using a consistent
framework which allows to draw specific conclusions but also to share data and support
transdisciplinary conclusions would be beneficial. Coupling material flow analysis with LCA should
be pursued in order to understand better the environmental impact of materials, such as energy
consumption studied by Yang and Kohler (2008), or global warming studied by Pauliuk et al. (2013)
and Pauliuk and Müller (2014).
6.2. Developing the conceptual framework for socioeconomic metabolism
analysis
Brunner (2011) identifies three phases of urban development for the definition of specific urban
mining strategies: growth (sometimes fast, as in Asia), steady state (sometimes temporary), and

shrinkage. These three phases or contexts can be observed in the reviewed studies. Fast growth can be
observed in studies dealing with China and Beijing; steady state (or reduced growth) in countries like
France, Germany, Japan, Norway, Switzerland or the USA; some cities in Saxony (Germany) are
shrinking. The consumption of construction materials, as well as the “maturity” of the stock and hence
the availability of secondary resources, differs for each of these contexts.
But the case study of Paris region also shows that within an urban area, different dynamics of
consumption can be observed (Barles, 2009, 2014). Similarly, Fishman et al. (2015) demonstrate that
stocks within Japanese prefectures were impacted differently by socio-economic drivers since 1945.
Thus, knowledge on socio-metabolic transition as defined by Fischer-Kowalski (2011) will be
improved by taking into account the social, economic, political and technical factors influencing the
mechanisms of the anthropogenic stock accumulation and removal within urban areas or
socioeconomic systems. Territorial ecology addresses these questions but remains insufficiently
developed (Barles, 2010).
Besides, ongoing research about Paris region
ii
shows that this area, which is already developed but still
experiences relatively significant urbanisation and population growth, is facing an emerging scarcity
of local resources (especially aggregates) due to land use conflicts between urbanisation and mineral
extraction. Whereas this situation is favourable for urban mining, relatively high building density and
a dynamic real estate market limit vacant space within the urban area that could be used for the
transformation and temporary storage of secondary resources. Hence spatial constraints limit local
extraction of primary and secondary resources within the region. This example points out that
prospective flows studies should increasingly consider the socio-spatial factors limiting present and
future use of secondary resources in order to contribute effectively to the implementation of industrial
ecology policies. Socioeconomic metabolism analysis calls for interdisciplinary research that would
associate engineering and social sciences.
6.3. Supporting industrial ecology policies
Anthropogenic flows and stock analysis can lead to the implementation of industrial ecology or
circular economy policies, as the case of Geneva illustrates. The initial work by Faist Emmenegger
and Frischknecht was followed by detailed studies about the factors limiting recycling (Rochat et al.,
2006). This led to various actions for the recycling of construction materials: production of a technical
guide about the use of recycled aggregates, communication to develop the awareness of professionals
for recycled products, development of recovery spaces, construction of experimental buildings and use
of recycled aggregates in government building projects (Erkman, 2005).
Other fields of action should be explored. The general framework for industrial ecology policies by
Erkman (1998) points out four fields: waste recycling; closing of material flows and minimisation of
dissipative emissions; dematerialisation of products and economic activities; and decarbonisation of
energy. Waste recycling may result from the development of local recovery industries (reuse and
recycling) as in Geneva, and it draws on incentives and coercive measures as R&D subsidies,
labelling, training, taxes, or adaptation of construction standards.
Secondly, closing of material flows and minimisation of dissipative emissions require a change in
building techniques, to enable the dismantling of building components and the reduction of losses
associated with potentially dissipated materials or with materials used in a dissipative way in
underground construction (Hashimoto et al., 2009). This could be achieved through buildings
specifications at a local scale or through a change in construction standards at a larger scale.
Thirdly, dematerialisation of products and economic activities may result from long term real estate
and networks management strategies that would aim to extend the lifetime of existing buildings and
networks through their conservation or transformation (Kohler and Hassler, 2002). It is also addressed
in planning and development policies as it "implies rethinking urban structures to minimise
infrastructure stock (roads, parking lots, etc.)" and "reducing the consumption of resources caused by
the structure and the sprawl of urban fabric (fuel for transport, water and energy networks for low-
density housing development)" (Erkman, 1998, p. 182, authors’ translation). However
dematerialisation is a tricky question in countries experiencing a high urbanisation rate as the material

stock is still being developed. Undergoing prospective work led by International Resource Panel
(UNEP) addresses this question (“Resource Requirements of Future Urbanization” project).
Finally, decarbonisation of energy raises questions about previous actions as regard to greenhouse gas
emissions generated by the production of construction materials, their transport and end-of-life
management. Linking material flow analysis to life-cycle analysis, as was carried out for instance on
the case of China by Yang and Kohler (2008), would bring a better understanding of the impacts of
material and energy flows on global warming. Decarbonisation is also addressed by policies for the
passive renovation of dwellings, whose impact could be better anticipated (Pauliuk et al., 2013;
Pauliuk and Müller, 2014).
In addition, the review shows that construction materials issues should be addressed by policies which
are usually considered as beyond the framework for industrial ecology policies. The relation between
construction materials consumption and urban forms calls for a better integration of this issue in urban
planning. This is also true at a larger spatial scale for land planning. Moreover, the review shows that
urban and territorial metabolic issues can’t be contained within the water-energy-food nexus. Indeed,
whereas flows (of water, food and energy) circulate through networks, inflows of construction
materials are also implied by their development and maintenance. Hence industrial ecology policies
dedicated to construction materials should be linked with other policies which impact those flows and
stock.
Conclusion
Flows and stock of building materials in cities like Beijing, Geneva, Orléans or Vienna, and countries
like China, Germany, Japan, the Netherlands, Norway, Switzerland or the USA have been estimated.
Material content, age, and location of secondary resources contained in the anthropogenic stock have
been studied. Retrospective studies have improved the knowledge on stock accumulation and removal
mechanisms and prospective modelling enabled partly to anticipate them. Research on flows and stock
of construction materials has led to the development of a methodological framework. It has also
allowed the conceptual framework for the analysis of socioeconomic systems to be completed, while
supporting the implementation of some industrial ecology policies.
Further research should address the issues that methods still raise in terms of assumptions and data,
notably material intensities and lifetimes. The development of case studies and the coupling of top-
down and bottom-up approaches would improve the reliability of estimates. Moreover, studies should
better take into account the social, economic, political and technical factors influencing the
mechanisms of stock accumulation and removal within urban areas, as well as the socio-spatial factors
limiting the use of secondary resources. Such research would contribute effectively to the
implementation of industrial ecology policies that would not only focus on recycling but also on the
closing of material flows and minimisation of dissipative emissions, the dematerialisation of products
and economic activities and the decarbonisation of energy and which should be linked with other
policies such as urban and land planning.
Acknowledgements
This research has been carried out within the framework of a project undertaken by the research centre
Géographie-Cités - CRIA (UMR CNRS 8504), and is financed by the Regional and Inter-departmental
Directorate for the Environment and Energy Ile-de-France (Direction régionale et interdépartementale
de l'environnement et de l'énergie Ile-de-France) and the Regional Council of Ile-de-France (Conseil
Régional d’Ile-de-France), which we thank for their support.

Bibliography
Publications reviewed
BARLES, S., 2009. "Urban metabolism of Paris and its region", Journal of Industrial Ecology 13(6):
898-913.
BARLES, S. 2014. "L’écologie territoriale et les enjeux de la dématérialisation des sociétés: l’apport
de l’analyse des flux de matières". Développement durable et territoires. Économie, géographie,
politique, droit, sociologie 5(1).
BERGSDAL, H., BRATTEBØ, H., BOHNE, R.A., MÜLLER, D. B. 2007. "Dynamic material flow
analysis for Norway’s dwelling stock", Building Research & Information 35(5): 557–570.
BRATTEBØ, H., BERGSDAL, H., SANDBERG, N. H., HAMMERVOLD, J., MÜLLER, D. B.
2009. "Exploring built environment stock metabolism and sustainability by systems analysis
approaches", Building Research & Information 37(5-6): 569-582.
DAXBECK, H., BUSCHMANN, H., NEUMAYER, S., BRANDT, B. 2009. Methodology for
mapping of physical stocks. Deliverable 2.3. Resource Management Agency (RMA). 77p.
DEILMANN, C. 2009. "Urban Metabolism and the Surface of the City", in: Guiding principles for
spatial development in Germany, German Annual of Spatial Research and Policy. Springer Berlin
Heidelberg. p. 98–112.
FAIST EMMENEGGER, M., FRISCHKNECHT, R. 2003. Métabolisme du canton de Genève. Phase
1. Rapport final pour le compte du groupe de travail interdépartemental Écosite de la République et
du canton de Genève. Uster: ESU service. 47 p.
FISHMAN, T., SCHANDL, H., TANIKAWA, H., WALKER, P., KRAUSMAN, F. 2014.
"Accounting for the material stock of nations", Journal of Industrial Ecology 18 (3): 407-420.
FISHMAN, T., SCHANDL, H., TANIKAWA, H. 2015. "The socio-economic drivers of material
stock accumulation in Japan’s prefectures", Ecological Economics 113: 76–84.
HASHIMOTO, S., TANIKAWA, H., MORIGUCHI, Y. 2007. "Where will large amounts of material
accumulated within the economy go? – A material flow analysis of construction minerals for Japan",
Waste Management 27: 1725-1738.
HASHIMOTO, S., TANIKAWA, H., MORIGUCHI, Y. 2009. "Framework for estimating potential
wastes and secondary resources accumulated within an economy – A case study of construction
minerals in Japan", Waste Management 29: 2859-2866.
HU, M., BERGSDAL, H., VAN DER VOET, E., HUPPES, G., MÜLLER, D. B. 2010a. "Dynamics of
the urban and rural housing stocks in China", Building Research & Information 38(3): 301–317.
HU, M., VAN DER VOET, E., HUPPES, G. 2010b. "Dynamic material flow analysis for strategic
construction and demolition waste management in Beijing", Journal of Industrial Ecology 14(3): 440–
456.
HUANG, T., SHI, F., TANIKAWA, H., FEI, J., HAN, J. 2013. "Materials demand and environmental
impact of buildings construction and demolition in China based on dynamic material flow analysis",
Resources, Conservation and Recycling 72: 91-101.
KAPUR, A., KEOLEIAN, G., KENDALL, A., KESLER, S. E. 2008. "Dynamic modeling of in‐use
cement stocks in the United States", Journal of Industrial Ecology 12(4): 539-556.
KOHLER, N., HASSLER, U. 2002. "The building stock as a research object", Building Research &
Information 30(4): 226-236.
KOHLER, N., YANG, W. 2007. "Long-term management of building stocks", Building Research &
Information 35(4): 351–362.

LICHTENSTEIGER, T., BACCINI, P., 2008. "Exploration of urban stocks", Environmental
Engineering and Management Journal 18(1): 41-48.
MÜLLER, D. B., BADER, H. P., BACCINI, P. 2004. "Long‐term Coordination of Timber Production
and Consumption Using a Dynamic Material and Energy Flow Analysis", Journal of Industrial
Ecology 8(3): 65-88.
MÜLLER, D. B. 2006. "Stock dynamics for forecasting material flows - Case study for housing in The
Netherlands", Ecological Economics 59: 142–156.
OBERNOSTERER, R., BRUNNER, P. H., DAXBECK, H., GAGAN, T., GLENCK, E., HENDRIKS,
C., … & REINER, I. 1998. Materials accounting as a tool for decision making in environmental
policy: MAcTEmPo case study report. Urban metabolism, the city of Vienna. Institute for Water
Quality and Waste Management, Vienna University.
ROUVREAU, L., MICHEL, P., SERRAND, M., MONTFORT CLIMENT, D., JAYR, E., PAPINOT.,
P.-E. 2012. Projet ANR ASURET - Analyse de flux de matière du secteur de la construction à l’échelle
de l’ouvrage et du territoire (tâche 4.2) - Rapport BRGM/RP-61849-FR. 142 p.
SANDBERG, N. H., SARTORI, I., BRATTEBØ, H. 2014a. "Using a dynamic segmented model to
examine future renovation activities in the Norwegian dwelling stock", Energy and Buildings 82: 287–
295.
SANDBERG, N. H., SARTORI, I., BRATTEBØ, H. 2014b. "Sensitivity analysis in long-term
dynamic building stock modeling - Exploring the importance of uncertainty of input parameters in
Norwegian segmented dwelling stock model", Energy and Buildings 85: 136–144.
SARTORI, I., BERGSDAL, H., MÜLLER, D. B., BRATTEBØ, H. 2008. "Towards modelling of
construction, renovation and demolition activities: Norway’s dwelling stock, 1900–2100", Building
Research & Information 36(5): 412–425.
SCHILLER, G. 2007. "Urban infrastructure: challenge for resource efficiency in the building stock",
Building Research & Information 35(4): 399-411.
SERRAND, M., MONFORT-CLIMENT, D., VAXELAIRE, S., MICHEL, P., GAUTIER, A.-L. 2013.
Projet ANR ASURET - Méthodologie d’aide à la décision : Evolution prospective des besoins en
matériaux et de la disponibilité des ressources primaires et secondaires - Exemple de la Ville
d’Orléans (tâche 5) - Rapport final BRGM/RP -62586-FR. 55 p.
SHI, F., HUANG, T., TANIKAWA, H., HAN, J., HASHIMOTO, S., MORIGUCHI, Y. 2012.
"Toward a low carbon-dematerialization society – Measuring the materials demand and CO2
emissions of building and transport infrastructure construction in China", Journal of Industrial
Ecology 16(4): 493–505.
TANIKAWA, H., FISHMAN, T., OKUOKA, K., SUGIMOTO, K. 2015. "The weight of society over
time and space: a comprehensive account of the construction material stock of Japan, 1945–2010",
Journal of Industrial Ecology 19 (5): 778–791.
WIEDENHOFER, D., STEINBERGER, J. K., EISENMENGER, N., HAAS, W. 2015. "Maintenance
and expansion: Modeling material stocks and flows for residential buildings and transportation
networks in the EU25", Journal of Industrial Ecology 19(4): 538–551.
YANG, W., KOHLER, N. 2008. "Simulation of the evolution of the Chinese building and
infrastructure stock", Building Research & Information 36(1): 1-19.
Other references
ADRIAANSE, A., BRINGEZU, S., HAMMOND, A., MORIGUCHI, Y., RODENBURG, E.,
ROGICH, D., SCHÜTZ, H. 1997. Resource flows: the material basis of industrial economies. 65 p.

AKSÖZEN, M., HASSLER, U., KOHLER, N. 2016. "Reconstitution of the dynamics of an urban
building stock", Building Research & Information: 1-20.
AYERS, R. U. 1994. "Industrial metabolism: theory and policy", The Greening of Industrial
Ecosystems, p. 23-27.
BACCINI, P., BADER, H.-P. 1996. "Regionaler Stoffhaushalt: Erfassung, Bewertung und Steuerung",
Heidelberg7 Spektrum Akademischer Verlag.
BARLES, S. 2009. "Urban metabolism of Paris and its region", Journal of Industrial Ecology 13(6):
898-913.
BARLES, S. 2010. "Society, Energy and Materials: What are the Contributions of Industrial Ecology,
Territorial Ecology and Urban Metabolism to Sustainable Urban Development Issues? ", Journal of
Environmental Planning and Management 53(4): 439-455.
BIRAT, J.P., DAIGO, I., MATSUNO, Y. 2013. "Methods to Evaluate Environmental Aspects of
Materials", p. 1459-1505, in : SEETHARAMAN, S. (ed.). Treatise on Process Metallurgy, Volume 3:
Industrial Processes, Newnes.
BRADLEY, P. E., KOHLER, N. 2007. "Methodology for the survival analysis of urban building
stocks", Building Research & Information 35(5): 529-542.
BRUNNER, P. H. 2011. "Urban Mining, A Contribution to Reindustrializing the City", Journal of
Industrial Ecology 15 (3): 339-341.
BRUNNER, P. H., RECHBERGER, H. 2004. Practical handbook of material flow analysis. Boca
Raton : Lewis Publishers. 336 p.
DOUGLAS, I., LAWSON, N. 2002. "Material flows due to mining and urbanization" in AYRES RU y
LW (eds.) 2002. Handbook of Industrial Ecology. p. 351-364.
ERKMAN, S. 1998. Vers une écologie industrielle : comment mettre en pratique le développement
durable dans une société hyper-industrielle. Paris, La librairie FPH. Edition Charles Léopold Mayer.
147 p.
ERKMAN. 2005. Ecologie industrielle à Genève. Premiers résultats et perspectives. GEDEC. 47 p.
EUROSTAT. 2001. Economy wide material flow accounts and balances with derived resource use
indicators. A methodological guide. Luxembourg: Office for Official Publications of the European
Communities. 85 p.
FISCHER-KOWALSKI, M. 2011. "Analyzing sustainability transitions as a shift between socio-
metabolic regimes", Environmental Innovation and Societal Transitions 1: 152–159.
GRÜN, A., BEHNISCH, M., KOHLER, N. 2009. "Perspectives in the reality-based generation, n D
modelling, and operation of buildings and building stocks", Building Research & Information 37 (5-
6): 503-519.
HSU, F.-C., ELVIDGE, C. D., MATSUNO, Y. 2013. "Exploring and estimating in-use steel stocks in
civil engineering and buildings from night-time lights", International Journal of Remote Sensing
34(2): 490–504.
KENNEDY, C., CUDDIHY, J., ENGEL-YAN, J. 2007. "The changing metabolism of cities", Journal
of Industrial Ecology 11(2): 43-59.
KLEEMANN, F., LEDERER, J., ASCHENBRENNER, P., RECHBERGER H., FELLNER, J. 2014.
"A method for determining buildings’ material composition prior to demolition", Building Research &
Information (ahead-of-print): 1-12.
KROOK, J., CARLSSON, A., EKLUND, M., FRANDEGARD, P., SVENSSON, N. 2011. "Urban
mining: hibernating copper stocks in local power grids", Journal of Cleaner Production 19: 1052-
1056.

LIANG, H., TANIKAWA, H., MATSUNO, Y., DONG, L. 2014. "Modeling in use steel stock in
China’s buildings and civil engineering infrastructure using time-series of DMSP/OLS night-time
lights", Remote Sensing 6(6): 4780–4800.
MATTHEWS, E., AMANN, C., BRINGEZU, S., FISCHER-KOWALSKI, M., HUTTLER, W.,
KLEIJN, R., MORIGUCHI, Y., OTTKE, C., RODENBURG, E., ROGICH, D. 2000. The weight of
nations: Material outflows from industrial economies.Washington, DC: World Resources Institute.
MLLER, E., HILTY, L. M., WIDMER, R., SCHLUEP, M., FAULSTICH, M. 2014. "Modeling
metal stocks and flows: A review of dynamic material flow analysis methods", Environmental science
& technology 48(4): 2102-2113.
ORTLEPP, R., GRUHLER, K., SCHILLER, G. 2015 "Material stocks in Germany's non-domestic
buildings: a new quantification method", Building Research & Information: 1-24.
PAULIUK, S., SJOSTRAND, K., MULLER, D. B. 2013. "Transforming the Norwegian dwelling
stock to reach the 2 degrees Celsius climate target", Journal of Industrial Ecology 17: 542–554.
PAULIUK, S., MULLER, D. B. 2014. "The role of in-use stocks in the social metabolism and in
climate change mitigation", Global Environmental Change 24(1): 132–142.
QUINN, D. J. 2012. Estimating material and energy intensities of urban areas. Doctoral dissertation.
Massachusetts Institute of Technology.
ROCHAT, D. ERKMAN, S., CHAMBAZ, D. 2006. Le recyclage des matériaux de construction à
Genève. 64 p.
STEADMAN, J. P. 1994. "Built forms and building types: some speculations", Environment and
Planning B: Planning and Design 21(7): S7-S30.
TANIKAWA, H., HASHIMOTO, S. 2009. "Urban stock over time: spatial material stock analysis
using 4d-GIS", Building Research & Information 37(5-6): 483-502.
WU, Z., ANN, T. W., SHEN, L., LIU, G. 2014. "Quantifying construction and demolition waste: An
analytical review", Waste Management 34(9): 1683-1692.
YOST, P., HALSTEAD, J. 1996. "A methodology for quantifying the volume of construction waste",
Waste Management & Research 14: 453–461.
i
The Construction History Society was founded in 1982, the international journal Construction History is
published since 1985, and the first international congress on construction history was held in 2003.
ii
Interviews with local stakeholders as part of a Ph.D. project about construction materials flows and stock in
Paris region.