Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
CLIMATECON Climatecon Working Paper Series
No. 1-‐2011
DEVELOPING A PRAGMATIC APPROACH TO ASSESS URBAN METABOLISM IN EUROPE A REPORT TO THE EUROPEAN ENVIRONMENT AGENCY
Dr. Jan Minx, Dr. Felix Creutzig, Verena Medinger, Tina Ziegler, Anne Owen and Dr. Giovanni Baiocchi
18/02/2011
Department of Climate Change Economics Technische Universität Berlin Room EB 238-‐240 (EB 4-‐1) -‐ Straße des 17. Juni 145, 10623 Berlin
www.climatecon.tu-‐berlin.de
2 | P a g e
Developing a Pragmatic Approach to Assess Urban Metabolism in Europe
Final Report to the European Environment Agency Project Reference: EEA/NSV/09/001 February 2011 Suggested citation: Minx, J.C., Creutzig, F., Medinger, V., Ziegler, T., Owen, A. and Baiocchi, G., 2011, Developing a Prgamatic Approach to Assess Urban Metabolism in Europe A Report to the Environment Agency prepared by Technische Universität Berlin and Stockholm Environment Institute, Climatecon Working Paper 01/2011, Technische Universität Berlin. Disclaimer The contents of this paper does not necessarily represent the official opinion of the European Environment Agency Ackowledgement: This report greatly benefited from the input of the following experts: Birgit Georgi, Mirko Gregor, Michael Narberhaus, Katy Roelich, Peter Christensen, Pawel Kazmierczyk, Jaume Fons, Christina Garzillo, Helga Weisz, Helge Brattebø, Maria Berrini, Branislav Olah, Michael Förster. Howeresponsibility for all errors made.
3 | P a g e
1 INTRODUCTION 5
2 URBAN METABOLISM -‐ INTRODUCING THE BASIC CONCEPT 7
3 TOWARDS AN EXTENDED CONCEPT OF URBAN METABOLISM 10
3.1 Extension 1: From environmental pressures towards aspects of environmental quality 10
3.2 Extension 2: Urban Drivers & Urban Patterns 12
3.3 Extension 3: Urban Quality & Co-‐Benefits 13
3.4 An extended concept for urban metabolism 13
4 TOWARDS A PRAGMATIC OPERATIONALISATION OF THE EXTENDED METABOLISM CONCEPT 15
4.1 Approach 1: A simple indicator system for monitoring urban metabolism in Europe 16 4.1.1 Urban Flow Indicators 19 4.1.2 Urban Drivers 26 4.1.3 Urban Patterns 32 4.1.4 Urban Quality 37 4.1.5 Headline indicator set 42 4.1.6 Applications 43 4.1.7 Towards an urban metabolism database 44
4.2 Approach 2: Small area estimates for carbon footprints and energy consumption 46 4.2.1 General introduction 46 4.2.2 Quick data validation attempt 53 4.2.3 Implications 55
5 DISCUSSION 56 5.1.1 Taking a systems approach 56 5.1.2 Linking to eco-‐system services and aspects of environmental quality 58 5.1.3 Urban drivers and patterns 60 5.1.4 Urban system and spatial resolution 61 5.1.5 Data availability & data sources 62 5.1.6 Comparability and uncertainties 63
7 SUMMARY AND RECOMMENDATIONS 65
8 LITERATURE 72
9 ANNEX A: SHORT LITERATURE REVIEW 80
9.1 Material flow analysis 80
5 | P a g e
1 Introduction Although urban areas cover only 4% of the area in Europe, they are home to almost 75% of the European population. As population and activity hotspots European cities (and urban areas) impose environmental pressures far beyond the borders of their own territory into their global Hinterland through the resource requirements of their production and consumption activities. Authors have therefore started to systematically quantify the physical inflows and outflows to urban systems and
(Wolman 1965). Urban areas are highly dynamic and accordingly their metabolism changes: with improving accessibility and stronger connectivity, urban development moves from single (sprawling) cities to a more disperse urban pattern across Europe and the formation of metropolitan areas. Urban areas increasingly use resources from abroad, impacting on areas far away, and thus become more and more dependent on remote areas influencing also their resilience. These factors, as well as demography and lifestyles, change the metabolism regarding intensities, distribution, dependencies and resilience.
This report has two overriding objectives:
The development of a conceptual framework to capture urban metabolism in Europe, which can adequately describe the functionalities, assess the environmental impacts of urban areas/patterns as well as ongoing urbanisation processes across Europe, show the inter-‐linkages and mutual impacts among urban areas and between urban and rural areas, and identify the drivers and successful response measures;
The provision of a first pragmatic approach to assess the environmental impact of urban areas and urbanisation processes from a European perspective and identify the role of different drivers.
To achieve the objectives, we progressed in four steps:
Based on a thorough literature review, we defined and agreed the conceptual framework for urban metabolism with the European Environment Agency (EEA) as a basis for assessing the
identifying the role of different drivers; We identified a relevant and feasible indicators describing the metabolism and consequent
impacts based on data review; We tested the approach based on a selection of representative European cities/urban areas
including cities of different population size and density and different regions of Europe having also a rather typical data situation;
Finally we derived recommendations for implementing this indicator framework from publicly available data sources, identified key data gaps as well as the most important areas for future research.
All these activities are taking place in the context of the Integrated Urban Monitoring in Europe (IUME) initiative started by the EEA. IUME is an attempt by the EEA to integrate the various urban monitoring initiatives across Europe with the ambition to identify and fill data gaps, improve the efficiency of work, and to provide an integrated information base and monitoring of progress towards more sustainable urban development. The IUME work is arranged in three basic components (see Figure 1):
6 | P a g e
The Data component is crucial to provide evidence and to quantify the urban development analysed at the appropriate scale. It aims at the identification of available data, data gaps, the links between different data sets and the appropriate tools to use them. The European Topic Centre on Land Use and Spatial Information (ETC-‐LUSI) has developed a framework for urban monitoring in Europe (Fons-‐Esteve et al. 2008) and initiated the data workstream of IUME.
The Question component defines research/policy questions regarding urban development and its likely impacts. IUME is designed to answer these questions. The EEA has set out an initial series of guiding questions in the context of its IUME activities.1
Understanding the urban system is crucial to reflect the interlinkages between the different drivers of urbanisation, arising pressures and impacts, and to identify appropriate response measures. It assists to develop the theoretical framework of the monitoring concept to link data and find answers to the complex policy questions.
Figure 1 -‐ The Integrated Urban Monitoring in Europe (IUME) approach
In this report we, therefore, start-‐framework for urban metabolism and a pragmatic approach for its assessment at the European level. Rather than deriving specific insights into the metabolism of individual cities and guide local policy action, it is the aim here to identify general trends in urbanisation across Europe and its underlying drivers and the evaluation of economic, social and environmental consequences. This aim will have implications for the choice of indicators, their specific definition and the type of analysis we undertake and recommend here. In this sense our research is different from
(Hertle and Schaechtele 2009) or 2
(Berrini and Bono 2007), TISSUE3, or early urban analysis undertaken by the EEA in its first Dobris Assessments (e.g. European Environment Agency 1996). However, this piece of work fundamentally differs from all attempts by trying to develop a pragmatic approach based on publicly available data. If successful, this raises hopes that a continuous monitoring becomes feasible at little costs that might provide a basic understanding of economic, social and environmental consequences of urbanisation in Europe and complement other more detailed assessments.
1 See, http://iume.ew.eea.europa.eu/about-‐1 2 See, http://status-‐tool.iclei.org/content.php/frontpage/?p=1 3 See, http://cic.vtt.fi/projects/tissue/index2.html
7 | P a g e
2 Urban Metabolism -‐ Introducing the basic concept The concept of Urban Metabolism goes back to Abel Wolman (see 1965), who was the first to draw the comparison between an organism and a city. Cities, like organisms, need energy and resources such as fuel, water or food as inputs to sustain life. These input are processed and ultimately released back to the environment as wastes. Hence, the basic rationale behind the urban metabolism concept is that the relationship between the environment and an urban system can be described by systematically recording all flows to and from the environment in physical terms as shown in Figure 2 in analogy to economy-‐wide material flow accounting (Eurostat 2001) or similar approaches (see Brunner and Rechberger 2004). In the absence of further information about environmental sources and sinks, this is then usually regarded as an estimate of the pressure environmental pressures generated by urban systems.
Figure 2 -‐ Basic metabolism concept: Physical exchanges between the urban system and the environment
Figure 3 identifies three basic types of metabolic flows that can be distinguished:
Direct Extraction and Releases: These are the resources directly extracted and the waste and emissions directly released within the urban system;
Imports and Exports: These are the products imported or exported to/from the urban system;
Indirect flows associated with imports and exports: These are the resources indirectly extracted and emissions and wastes indirectly released in the supply chain of goods and services imported to or exported from the urban system.
Note that only metabolic flows are recorded, which cross the boundary between the environment and the urban system. Capital accumulation in physical terms is therefore usually only represented in terms of net additions to stocks initially. However, a more comprehensive stock description is desirable for understanding urban metabolism for at least two reasons:
Stock accumulation causes outflows delays in the urban metabolism. This issue is particularly important when long-‐term quantitative estimates are of concern (e.g. scenario analysis);
Stock characteristics such as the size, age or energy efficiency of a residential building can pre-‐determine a considerable share of the metabolic flows of a city.
8 | P a g e
Figure 3 -‐ Basic metabolism concept: Categories of metabolic metabolic in-‐ and outflows
As suggested by Figure 2 and Figure 3 the urban metabolism concept is based on the idea that environmental pressures generated by urban life need to be assessed in a systems approach. All metabolic inflows and outflows to the urban system have to be quantified. Such a systems approach has two main features:
Completeness in the description of metabolic flows: A complete description of all metabolic flows is important to detect environmental problem shifting associated with policies. For example, we might reduce carbon dioxide emissions by building nuclear power plants, but society has to deal with the resulting waste for centuries to come. In contrast, CO2 emission reductions from energy efficiency improvements might have very little environmental side effects. Only by tracking all material interactions between the environment and the urban
can be avoided or minimised in the long-‐run.
Global system boundaries and consumption-‐based accounting: Urban areas are hotspots of human life with high concentrations of people and activities. Due to these space limitations cities and urban areas are often highly reliant on their (regional and global) hinterland. A considerable share of environmental pressures associated with urban life are generated elsewhere in the world and imported to cities. This could be any place in the world and we are therefore confronted with global system boundaries when we deal with the concept of urban metabolism. Hence, the urban metabolism concept requires taking into account the
are generated in the global supply chain of goods and services. Consumption based accounting and a complete description of upstream environmental pressures are therefore the second systemic feature of the urban metabolism concept. A definition of consumption and production based accounts is given in Box 1.
These two features of the urban metabolism approach pose very high data requirements. In fact, in the literature there are no studies, which provide a comprehensive quantification of urban metabolism. However, different sets of studies have been able to address different aspects. Material flow analysis studies of cities (e.g. Hendriks et al. 2000; Warren-‐Rhodes and Koenig 2001; Browne et
9 | P a g e
al. 2009; Niza et al. 2009), for example, have focussed on trying to capture a large variety of metabolic flow types (e.g. water, fossil fuels, sewage solid waste etc.). Due to the data intensity of the task and severe restrictions in terms of data availability, these studies only quantify the materials directly imported/exported to/from a city. The indirect material requirements in higher supply chain layers are neglected. In contrast, there is an increasing number of studies focussing on specific metabolic flows such as energy or CO2 emission flows associated with cities. These studies are often able to quantify all the indirect energy and CO2 requirements associated with cities (e.g. Druckman et al. 2008; Ramaswami et al. 2008; Kennedy et al. 2009; Minx et al. 2009; Hillman and Ramaswami 2010; Kennedy et al. 2010). A more comprehensive literature review has been published in earlier project reports: a summary is provided in Annex A.
Hence, while the urban metabolism concept as a systems approach establishes wide system boundaries on a conceptual level, limited data availabilities force focussing on specific aspects of the metabolism when it comes to practical implementations. However, also on the conceptual level this basic metabolism concept introduced above can be critiqued. Based on such criticism, three conceptual extension of the basic metabolism concept will be introduced in the next Section.
There are two fundamental types of physical accounts. Production based accounts comprise all material extraction and residual releases that occur on the territory of a city directly (or region, country etc.) regardless whether the activities triggering these flows serve the residents of the city (i.e. domestic activities/ domestic consumption) or people elsewhere in the world (i.e. foreign activities/ exports). In production based accounts the focus therefore is on all physical exchange processes taking place on the territory of a particular city for any beneficiary (city residents or anybody else). Consumption based accounts comprise all material extraction and residual release required for consumption in a particular city (or region, country etc.) directly or indirectly, wherever these physical flows might occur in the world. Compared to production based accounts they therefore include physical flows associated with imports and exclude physical flows associated with exports. Moreover, they provide a complete description of all direct and indirect metabolic flows required for a specific consumption. In consumption based accounts the focus is therefore on all physical exchange processes taking place anywhere world for a particular beneficiary (i.e. city residents).
Box 1 -‐ Production and consumption based accounts in the context of urban metabolism
10 | P a g e
3 Towards an extended concept of urban metabolism What has been presented so far can be regarded as the standard urban metabolism concept. It is the basis of most of the available evidence so far. However, we argue here that this basic metabolism concept needs to be extended. The extensions can be motivated by the particular requirements of this project such as:
the interest in moving from environmental pressures towards aspects of environmental quality and in linking the urban metabolism to eco-‐system service provision and eco-‐system functioning;
the interest in environmental pressures generated by urbanisation processes and urban sprawl;
Beyond these project requirements, there is a more general need for such extensions, because many studies available have remained at purely describing the metabolic inflows and outflows. However, unless we know how specific determinants such as urban form, lifestyles or the available infrastructure manifest in metabolic differences across cities and other urban settlements, the knowledge about size and types of metabolic in-‐ and outflows is of very limited use for understanding urban systems and informing local decision making processes. For similar reasons there is a need to link/ juxtapose changes in the physical metabolism to/with changes in aspects of urban quality of life. In this Section we will therefore briefly propose three relevant extensions to the urban metabolism concept before we enter discussions on how such an extended concept can be operationalised. The Section is based on ideas and terminology mainly developed by Marina Alberti (Alberti 1996; Alberti 1999; Alberti et al. 2003; Alberti 2005).
3.1 Extension 1: From environmental pressures towards aspects of environmental quality
One shortcoming of the standard urban metabolism concept is that it only provides information on environmental pressures in terms of the amount of resources extracted or the amount of pollution generated. Little information is usually provided in terms of how this might change aspects of environmental quality or how this might relate to basic concepts of environmental sustainability such as resilience or carrying capacity (e.g., Holling 1977). The literature on urban Ecological Footprints (e.g. Rees and Wackernagel 1996; Wackernagel 1998; Newman 2006) provides one notable exception. These studies link comprehensive material and product flow accounts to an inverse carrying capacity concept. Moreover, the Urban Ecology literature usually looks at specific aspects of ecosystem health or functioning in relationship to development of particular urban systems even though often fail to establish a link to the wider metabolism of cities. A good review is provided in Alberti (1999).
In order to move conceptually from environmental pressures towards aspects of environmental quality, we extend the urban metabolism concept in Figure 4 by adding three components explicitly into our framework: (1) environmental sources, (2) environmental sinks and (3) ecological support functions. Sources and sinks are already (implicit) dimensions of the standard urban metabolism concept as shown in Figure 2. However, both need to be further specification.
On the source side links need to be established between trends in urban resource use and the state of the natural resource base. This should take into account both the state of natural resources and
11 | P a g e
the biological processes that sustain them. Crucially, resource flows not only matter from an absolute perspective, but also in relation to overall resource availability. For example, the impact of consumption of (regional) wood depends on the current state of (regional) forest feedstock. The current state is a function of the historical resource consumption. Similarly, there is the need to establish a link between emission and waste releases and the capacity of the local, regional and global environment to absorb these. The impact of waste is not only a function of its absolute scale but depends on the absorptive capacity of the environment. The current capacity is a function of past sink reliance. Changes in both environmental sources and sinks affect the ability of the environment to provide life supporting services such as nutrient cycling, water purification or biological productivity or a viable climate. Current resource state and sink capacity not only influence nowadays ecological functions but will have also impact on the its future functioning. For example, greenhouse gas emissions will stay in the atmosphere for varying time scales. Altogether, as the first conceptual extension of the urban metabolism concept we propose to introduce state of resources, capacities of sinks, and ecological supporting functions to sources and sinks, and make explicit the relevant time scales.
Figure 4 -‐ Extending the urban metabolism concept for environmental impacts
While there seems little doubt about the value of extending the urban metabolism concept to aspects of environmental quality, the operationalisation is a major challenge. Establishing links between metabolic flows, environmental sources and sinks as well as ecosystem functioning is highly complex. However, the complexity is multiplied manifold by the systemic nature of the urban metabolism concept itself. This essentially means that a cities metabolism can be connected not only to global, but also to local and regional environmental problems anywhere across the globe, which are heavily dependent on local circumstances. Not only is it difficult to trace the metabolic flows back to particular locations through global supply chains, but it is equally difficult to assess how changes in the metabolism affect complex and dynamic ecosystems.
Therefore, the resulting challenge is not so much the extension of the conceptualisation of urban metabolism than finding feasible way for an operationalisation: Is it possible to describe some of these linkages to provide at least some basic insights into the changing environmental impacts triggered by urban life? What are the uncertainties involved? How do uncertainties trade-‐off with transaction costs for generating adequate evidence? It might not be possible to fully answer this
12 | P a g e
question in the context of this small tender. However, it might be possible to develop a general approach and identify avenues, which might be worthwhile pursuing in the future.
3.2 Extension 2: Urban Drivers & Urban Patterns Behind the interest in urban metabolism research lies the assumption that cities have a distinct metabolism, i.e. that size and type of metabolic flows of an urban area is influenced by its land-‐use patterns such as the its form, land-‐use intensity, population density but also its size. Metabolic flows are equally shaped by drivers such as land-‐use planning and infrastructure decisions or the economic role of the city under consideration as well as the lifestyles of its residents. All these sets of determinants are interdependent as shown in Figure 5. We will henceforth refer to these as urban drivers, urban patterns and urban lifestyles.4 Hence, we argue that one of the key questions for urban metabolism research is how trends in urban metabolic flows are linked to trends in spatial structure, urban organizations and lifestyles. Unless the urban metabolism concept addresses the relationship between urban drivers, urban patterns and urban drivers with urban metabolic flows, little can be learned from urban metabolism studies. We propose the inclusion of urban drivers, urban patterns and urban lifestyles as the second extension of the basic urban metabolism concept.
Figure 5 -‐ Urban drivers, urban patterns & urban lifestyles as determinants of urban metabolism
Urban metabolism studies and their underlying conceptual framework have given relatively little attention to these aspects so far. Even after several decades of research, there is still relatively little evidence how urban patterns, urban drivers and urban lifestyles change the metabolism of cities and their environmental impacts. Clearly, some research areas are better understood than others. For example, while there is little evidence on how urban infrastructures determine the lifestyles of a
energy use (and CO2 emissions) and urban patterns (Newman and Kenworthy 1989; Newman and Kenworthy 1996). By integrating urban patterns, urban lifestyles and urban drivers explicitly into an extended urban metabolism concept, we highlight the fundamental importance of these aspects for
4 In later stages it will be helpful for matters of simplification to consider urban lifestyles as one set of urban drivers, but we keep them separate for now in order to highlight their individual importance.
13 | P a g e
3.3 Extension 3: Urban Quality & Co-‐Benefits The final extension to the standard urban metabolism concept is the introduction of aspects of urban quality. We define urban quality broadly as local quality of life. This covers a variety of issues such as local environmental quality, human health, accessibility, employment opportunities, urban design quality and so on. As national policies ultimately aim to secure or improve the quality of life
often focus on measures that improve local quality of life. While local quality of life considerations will often be central to public perceptions of local policies, the metabolism concept is crucial to evaluate the global, system-‐wide consequences of these policies. However, depending on the local policy option chosen a metabolic change can improve, deteriorate or leave urban quality unaltered.
For example, to meet increased urban energy demands, a city can build a new power plant increasing local air pollution. This might have detrimental health effects for at least some of the urban population and therefore diminish quality of life in the city. Alternatively, it can import the additional energy from elsewhere. In this case, whilst the metabolism of the city grows, urban quality would not be affected in the city of consideration (but potentially elsewhere). However, the global environmental effects could potentially be negative (even though not necessarily). A third alternative could be the replacement of an old, small inefficient power plant with a new, bigger and highly efficient one , which provides more usable energy with a smaller pollution output. In this case depending on the degree of efficiency improvements and the level of additional energy demands, the metabolic change could even decrease local air pollution and improve urban quality.
This idea is not new. Along similar lines, authors have already suggested to expand the metabolism (Newman et al. 1996; Newman 1999; Timmer and
Seymoar 2005; European Environment Agency 2009). More prominently, in the climate change literature the idea that local air pollution policies could have substantial side-‐effects in terms of greenhouse gas emission savings is well discussed under the notion of `co-‐ (Aunan et al. 2004; Aunan et al. 2006) and it is widely accepted that co-‐benefits are a core component of local climate change policy particularly in developing countries. Similar extension to all system-‐wide metabolic flows could be seen as a generalisation of the co-‐benefits approach. Establishing a link between urban metabolic flows and aspects of urban quality is therefore proposed here as a third indispensable extension of the standard urban metabolism concept relevant for choosing the most appropriate policies and making conscious trade-‐offs between local and system-‐wide consequences.
3.4 An extended concept for urban metabolism The proposed, extended conceptual framework for urban metabolism is summarised in Figure 6 integrating the three proposed extensions. The metabolic inflows and outflows are its central dimension, but urban quality, the biophysical processes determining the environmental impacts associated with environmental sources and sinks as well as urban drivers, patterns and lifestyles are conceptually interlinked in a holistic approach. The framework distinguishes between three types of flows as previously highlighted in Figure 3: (1) direct extraction & releases; (2) imports and exports; (3) indirect flows associated with imports and exports. While our initial focus here will be on the physical aspects of the metabolism it is important to highlight that the 3-‐type distinction can be easily adjusted to also cover economic and social inflows and outflows such as information, cultural goods or employment.
15 | P a g e
4 Towards a pragmatic operationalisation of the extended metabolism concept
In this Section we move towards the quantification of the conceptual framework developed above. We first emphasise the requirements for the quantification of urban metabolism outlined in the context of this project. In a second step we introduce a three tiered research approach for this scoping study and end this Section by providing some details for the practical implementation in the next phase of the project.
It is the intention of the European Environment Agency (EEA) to apply the metabolism concept to
ongoing urbanisation processes across Europe, show the inter-‐linkages and mutual impacts among urban areas and between urban and rural areas, and identify the drivers and successful response
for a framework for urban metabolism:
Consumption based accounting: A framework for urban metabolism should be centred around a set of consumption based indicator in order to evaluate the global requirements on environmental sources and sinks triggered by urban final demands. This requires a systems approach similar to the one applied in life cycle analysis.
Completeness: A framework for urban metabolism should cover all types of metabolic inflows and describe them throughout the global supply chains of goods and services consumed in urban areas.
Beyond environmental pressures: A framework for urban metabolism should aim to go beyond environmental pressures and establish links to potential local, regional and global environmental impacts.
Interlinkages: A framework for urban metabolism should describe the relationship between metabolic inflows and outflows on the one hand, and urban flows, urban patterns and urban quality on the other hand, in order to be able to determine environmental impacts associated with on-‐going urbanization processes.
Pragmatism: The focus of the implementation of the conceptual framework is on pragmatism and therefore what can be done with existing information in the short term. The same pragmatism is applied when propositions are made how indicator framework can be improved in the future.
Comparability: Given the basic goal of understanding environmental impacts of wider urbanization processes across Europe, comparability of information is paramount.
Transparency: Given that local data quality might fluctuate considerably, high levels of transparency in terms of data sources and estimation methodology are required.
Human settlements: Given that urbanisation processes, urban sprawl and their environmental impacts emerge at the interface between rural and urban living, the urban metabolism framework should be applicable not only to urban areas but any human settlement.
Ideally we would have all the required data available at the appropriate spatial scale for operationalising the extended urban metabolism concept in a useful indicator framework. However, the real world situation is quite different. Even though there is a considerable amount of information
16 | P a g e
available at the city-‐scale level, data availability is fragmented, data are of different type and refer to different delineations of the urban system (Fons-‐Esteve et al. 2008). At the same time considerable data gaps exist, which are particularly severe when it comes to metabolic flow descriptions of the urban system, while this data situation can vary considerably across countries.
To distinguish available options that can be implemented today, opportunities arising with more detailed datasets in the future and to identify research avenues that should be pursued, we use a three tiered strategies in the empirical Section of this report focussing on quantifying urban metabolism. First, we will build a pragmatic, feasible indicator system for quantifying urban metabolism. This indicator system will use the administrative delineations of cities as boundaries,
patterns and urban quality. We will trial this indicator set for different European cities. The individual indicators will refer to cities as a single spatial entity. The data will be collected and validated.
Figure 7 -‐ Approach taken for empirical part of project
Second, we will address the need for data with a higher spatial resolution and robust methods for downscaling environmental information by scoping the potential of a geo-‐demographic approach to the quantification of urban metabolism. Geodemographic data sets (Harris et al. 2005) are much richer in terms of the number of variables contained and each variable is provided at very high levels
context of the neighbourhood they live in. This makes such data systems interesting for in depth analysis of the influence of urban flows, urban patterns and lifestyles on urban metabolic flows. We will use an existing UK dataset for evaluating the potentials of an europe-‐wide application of such systems. The third avenue will identify interesting areas for future research, which could not be covered in this scoping study.
4.1 Approach 1: A simple indicator system for monitoring urban metabolism in Europe
The purpose of an indicator system should dictate its design. In this report we want to understand the physical metabolism of European cities and its local, regional and global environmental consequences. Many existing indicator sets have been designed to inform and guide local policy. Our aim is the identification of a set of general determinants behind commonalities and differences in
17 | P a g e
the metabolism of cities (urban flows) across Europe and their relationship to urban structures (urban patterns), socio-‐economic drivers (urban drivers) and aspects of quality of life (urban quality). We will therefore construct indicators in a way that facilitates such an analysis. For example, while performance indicators measuring changes in a variable over time (against a standard, goal or reference value) are often useful to guide local decision-‐making, our ambitions will often require descriptive indicators, which describe the state of the urban system with respect to a particular attribute (e.g. amount of energy used, number of cars registered etc.). In the indicator design we further take into account data availabilities as well as relevant European policy agendas such as the Thematic Strategy on the Sustainable Use of Resources or the Leipzig Charta as well as European-‐wide local government initiatives such as the Aalborg Commitments.
Figure 8 shows the basic structure of the proposed indicator system. Consistent with the conceptual framework introduced above, it monitors metabolic inputs and outputs (urban flows) in the context of urban drivers, urban patterns and urban quality. Even though we will discuss indicators and data
will remain at the centre of attention here. Urban drivers, pattern and quality describe the conditions under which metabolic flows arise and provide the required contextual reference frame. Instead of devising a completely new indicator system we built on existing work carried out for the European Environment Agency for the first and second Dobris Assessment by Marina Alberti (Alberti 1996; European Environment Agency 1996) and more recent initiatives such as TISSUE (see Footnote 3) or Urban Ecosystem Europe (Berrini and Bono 2007).
Figure 8 -‐ Structuring a simple indicator system for monitoring urban metabolism (adapted from Alberti (1996))
Figure 9 further specifies the indicator framework by identifying thematic areas across the four dimensions of the proposed indicator systems (urban flows, drivers, patterns and quality). Choosing such thematic areas is always subjective to some extent. However, the choices made here are informed by a detailed analysis of data availabilities, a review of relevant academic literature (see Minx et al. 2009) as well the consideration of relevant policy documents.
18 | P a g e
Figure 9 -‐ A pragmatic indicator framework for quantifying urban metabolism in Europe (adapted from (Alberti 1996)
We have embedded the proposed indicator framework for quantifying urban metabolism in Europe into a four level information pyramid as shown in Figure 10. The basic data sources sit at the bottom
Figure 1). From these various data sources we propose the construction of an urban metabolism database at the next level. This database contains a wide range of information on the physical metabolism of cities in Europe. Some of this data might be readily extractable and other might need to be derived or imputed from existing databases. Note that this urban metabolism database focuses on physical flows only (and therefore mainly, but not exclusively on the urban flow dimension as some of the relevant regional and local pollutants are used as indicators in the urban quality dimension) and serves for (1) a more systematic and comprehensive description of the urban metabolism; (2) facilitating future research/ analysis; (3) enabling a systematic filling of data gaps. From the existing urban and GIS databases as well as the urban metabolism database we will derive our indicator set for quantifying urban metabolism in Europe. As this indicator set still contains a rather large amount of data, key information are summarised in a headline indicator set.
Figure 10 The indicator framework for quantifying urban metabolism in Europe as a four level information pyramid
In the next Sections we will introduce the proposed pragmatic indicator set starting with urban flows and subsequently moving to urban drivers, urban patterns and urban quality respectively. Afterwards we will summarise key indicators across these four dimensions in a headline indicator set and outline crucial analytical extensions in order to capture some fundamental relationships for the assessment of links between urbanisation and the metabolism of European cities. Finally, we will discuss how an urban metabolism database might look like (see Figure 10).
19 | P a g e
4.1.1 Urban Flow Indicators Urban flow indicators represent the physical metabolism of a city. Social aspects of the urban systems are covered in the urban driver and urban quality categories to some extent. In an ideal world we would be able to monitor the complete metabolism of a city and link to the various related social, economic and bio-‐physical processes around the world. However, for many of these aspects there is simply no data available. Accepting the requirement of a pragmatic approach in the indicator development based on publicly available data sources, we use the available urban audit and IUME data as a starting point and propose four thematic areas in the urban flow indicator dimension:
Energy & Climate Change; Water; Waste; Land-‐use;
A more systematic description of the metabolic inflows and outflows across the four thematic areas is shown in Figure 11. 5 There is no doubt that the selection of these metabolic flows is to some extent arbitrary. However, apart from arguing in terms of pragmatism and data availabilities, there is also a political justification. Since the publication of the Lund Declaration in 2009 the European Commission has been putting an emphasis on the necessity for the research and policy community to respond directly to a series of Grand Societal Challenges. Among the various challenges identified
inflows and outflows. These are largely covered in our pragmatic implementation. Only metabolic flows associated with food production and consumption in the urban system are only indirectly tackled (via land use indicators and global greenhouse gas emissions from food consumption).
Figure 11 -‐ Systematic metabolic description of inflows and outflows across the four thematic areas
5 The classification of land solely as an input is ambiguous, but serves the simplified depiction here.
20 | P a g e
Further note that, at a conceptual level, we address stocks partly in the urban flows (e.g. availability of renewable energy technologies) and partially in the urban patterns dimension of the indicator system (e.g. building stock, transport infrastructure etc.). However, these stock descriptions remain at a very general level and do not provide an indication on, for example, individual material stocks that have accumulated in urban infrastructures (see Rauch 2009).
The above figure provides a simplified diagram of the urban water cycle. Water is extracted within or outside the administrative boundaries of a city, where it is treated and distributed before it is used as part of different activities (A,B,C,D). The metabolism of a city in this context is determined by the water service level as well as other urban driver and patterns such as the socio-‐economic make-‐up of the population, the water infrastructure in place (leakage) etc.. After use the water is collected and treated again before it is released back to nature (or reuse). According to this scheme we can develop indicators for each major step of this chain. Water extraction indicators (step 1) should ideally provide an indication about the availability of water (water scarcity indicators), the share of domestic and imported water resources (location of major supplies) as well as source types. Water treatment and distribution (step 2) could be characterised by measures of water leakage and treatment technologies. Water use (step 3) should be distinguished by type and insufficient services levels. Most relevant aspects of waste water collection are the quantity of water collected, the share in stormwater overflows including related pollution as well as the share of waste water exports. Finally urban wastewater treatment (step 5) should provide information about the quantity treated, the share of treatment types and the pollution levels of wastewater treatment plant discharges. These indicators should be complemented and ideally integrated with information on the state of the water resources in the city as well as the providing regions. Potential indicators include pollution levels in water, drinking water quality or polluted drinking water abstractions from wells or groundwater.
Box 2 -‐ The Urban Water Cycle and related indicators
Table 1 provides an idealised indicator system of urban flow indicators. Each indicator is related to one of the three eco-‐system service categories introduced earlier. In the areas of water, energy and land this system juxtaposes the resource use and emission releases on the urban territory with consumption based indicators representing the global environmental resource demands or pollution releases. In the design of the indicators we put an emphasis on the productivity in the use of resources (European Commission 2005). This is reflected by expressing indicators relating to environmental sources in
21 | P a g e
terms of efficiencies and indicators relating to environmental sinks in terms of intensities. In general the development of indicators in each thematic area requires a thorough conceptualisation. We have exemplified this here for the area of water in Box 2. For a comprehensive description and a sound understanding of metabolic flows in each of these thematic areas a whole range of indicators could be devised. Here we have chosen those indicators, which we perceive to be most important for understanding urban metabolism across European cities. We propose the development of a more comprehensive database in a later Section, which can be continuously developed and contains a larger amount of data for in-‐depth studies and future extensions of the framework. We have labelled the indicator system in Table 1 as desirable due to the unavailability of some of the indicators from public, continuously updated databases. As wish list we have not described the indicators in detail, but remained with a general description.
In Table 2 we list relevant urban flow data, which can be compiled within the next six to twelve month. The main bottleneck within this system remains the energy and climate theme. Even though it has been attempted to collect this data as part of the Urban Audit, response rates were so low that the variables had to be dropped from the published dataset in the end due to low data availability (Brandmueller, 2010: personal communication).
Currently, the data would therefore have to be collected from the city authorities.6 For our testbed of five cities this turned out to be reasonably resource intensive (on both ends). Data was available for four of the five cities as shown in Table 3. Comparability of this data was an issue for two reasons: first, not all data provided corresponded to the requested sector breakdowns. Second, it remained difficult to understand how the emission inventories were compiled.
A variety of initiatives have emerged recently focussing on the construction of urban energy and emission inventories. These are led by different institutions such as the Covenant of Mayors, ICLEI, Climate Alliance, Eurocities, 2 degree initiative etc.. Some of these institutions are also trying to push forward a standardisation process for energy and emission inventory compilation at the local level (e.g. ICLEI 2009; Covenant of Mayors 2010). Even though it is unclear at this moment whether and to what extent the data compiled in under these various initiatives will be made publicly available (and how easy access would be), a variety of positive externalities are to be expected:
-‐ Increased data availability: For the next urban audit round, there should be considerably more energy and greenhouse gas emission data be available at the city level.
-‐ Increased comparability: A greater number of emission inventories will have used similar calculation methodologies due to the various standardisation efforts.
-‐ Increased transparency: It will be easier to understand how emissions have been calculated given that an increasing amount of emission data will follow certain standardised methodologies.
Availability of data in the areas of (solid) waste, water and land-‐use is less challenging acknowledging the need for a pragmatic data strategy. Even though it might not always be possible to construct the most desirable indicator from the available, the available data provides a reasonable starting point. The solid waste data can be fully sourced from the urban audit as well as some of the water data.
6 In fact, another way would be to gather the data from literature and past projects (LINK TISSUE)
22 | P a g e
Information on water scarcity and quality are very limited and very simplified indicators are used currently. Complementary water indicators particularly concerning the protection of the quality and supply of fresh water resources, water scarcity and leakage should be constructed from WISE7 or similar databases in the future. Land-‐use data can be taken from the Urban Atlas8 project, where data has already been compiled for larger urban zones with more than 100000 inhabitants as defined for the urban audit. CORINE provides complete coverage across Europe with less resolution and could be used to complement Urban Atlas, if required.
In terms of urban audit data one question is whether relevant variables are collected. The other question is the response rate of cities for particular variables. These are shown in Table 3 as well. For three of the five urban audit variables (water consumption, water infrastructure and water quality) data is available for all cities for at least one year. However, only for one of the three variables (water consumption) data is available for the same year across the cities. This is a general limitation of urban audit data. For two urban audit variables (waste collection, waste composition) data is only available for four of the five cities (no data for Lille). Across all urban audit cities data availability ranges between 35% and 65% depending on the urban flow variable under consideration.
In terms of the spatial delineation data of the available urban flow data, there is a strong tendency towards exclusive data availability for administrative areas. The problems associated with administrative delineations for urban research have been extensively discussed elsewhere (Fons-‐Esteve et al. 2008). However, it seems to be a reality that has to be accepted and dealt with: environmental city level statistics are sparse and if they are collected this commonly takes place at the city level. We will deal with this issue in two ways here: first, in the context of this indicator system we will try to develop indicators that shed some light into the physical make-‐up of the city territory in terms of its administrative boundaries. Second, we will discuss ways to downscale information in later sections of this report.
7 http://water.europa.eu/en/welcome 8 http://www.eea.europa.eu/data-‐and-‐maps/data/urban-‐atlas
23 | P a g e
No Area Name Description Ecosystem Source
Sink Functioning
1
Energy & clim
ate
CO2 intensity of production Annual direct CO2 emissions released from urban territory emitted by industry per unit of local GDP X 2 CO2 intensity of transportation Annual direct CO2 emissions of road transport sector per capita X 3 CO2 intensity of residential users Annual direct CO2 emissions of residential sector per capita X 4 Carbon Footprint Annual direct and indirect CO2 emissions from final consumption activities per capita X 5 Energy efficiency of production Annual energy use by industrial sector per unit of local GDP X 6 Energy efficiency of transportation Annual energy use of road transport sector per capita X 7 Energy efficiency of residential usage Annual energy use by residential sector per capita X 8 Renewable energy production Share of renewable sources produced on urban territory X 9 Energy footprint Annual direct and indirect energy use from final consumption activities per capita X 10
Water
Territorial water extraction Share of water extracted on urban territory X 11 Groundwater levels Change in groundwater level on urban territory over the last 5 years X (X) 12 Water scarcity An adequate indicator of water scarcity X (X) 13 Water use efficiency Annual amount of water used on urban territory per capita X 14 Waste water treatment Share of waste water released back into environment untreated X (X) 15 Water quality extraction Water quality of water extracted for urban use X 16 Water quality release Water quality of water released back into the environment X 17 Water footprint Annual amount of direct and indirect water use from final consumption activities in a city by major
categories (food, housing, transport, other) and region of water extraction X
18
Waste
Waste intensity of production Annual amount of solid waste collected from industrial sector per unit of local GDP X
19 Residential waste intensity Annual amount of solid waste collected from residential sector per capita X
20 Waste recycling Share of solid waste recycled X 21 Waste incineration Share of solid waste incinerated X
22 Landfill Share of solid waste landfilled X 23
Land
-‐use
Soil sealing Increase in soil sealing on urban territory by type of converted land X 24 Land Footprint Annual size of land directly and indirectly used in the production of goods and services finally
X
Table 1 -‐ A pragmatic but not yet feasible indicator system for monitoring urban metabolism across Europe
24 | P a g e
Table 2 -‐ Overview proposed urban flow data (C=cities; P=public; R=restricted)
Area Name Description Data Comments, Alternative data sources Source Spatial
Unit Data availability Continuit
y Case studies
other
UF1
Energy & Clim
ate
CO2 intensity of production
Annual direct industrial CO2 emissions released from urban territory per unit of local GDP
Cities C R Unknown
Varied Covenant of Mayors, Future urban audit, ICLEI
UF 2
CO2 intensity of transportation
Annual direct CO2 emissions of road transport per capita Cities C R Unknown
Varied Covenant of Mayors, Future urban audit, ICLEI
UF3
CO2 intensity of residential activities
Annual direct CO2 emissions of residential sector per capita Cities C R Unknown
Varied Covenant of Mayors, Future urban audit, ICLEI
UF4
CO2 intensity of energy use
CO2 emissions from energy use within and outside the city territory per capita
Various C R Unknown
Varied Various data sources; the development of the account should be embedded in a clear research agenda
UF5
Energy efficiency of production
Annual energy use by industrial sector per unit of local GDP Cities C R Unknown
Varied Covenant of Mayors, Future urban audit, ICLEI
UF6
Energy efficiency of transportation
Annual energy use of road transport sector per capita Cities C R Unknown
Varied Covenant of Mayors, Future urban audit, ICLEI
UF7
Energy efficiency of residential usage
Annual energy use by residential sector per capita Cities C R Unknown
Varied Covenant of Mayors, Future urban audit, ICLEI
UF8
Water
Territorial water extraction
Share of water resources extracted and used on urban territory Cities C P Unknown
Yes, 3 years
Water exploitation index is an alternative indicator; WISE database should be considered for future developments
UF9
Water use efficiency Annual amount of water used on urban territory per capita Urban Audit
C P 54%-‐67%
Yes, 3 years
Ideally an indicator such as water exploitation index
UF10
Waste water treatment
Share of dwellings connected to sewage system Urban Audit
C P 55%-‐57%
Yes, 3 years
UF11
Water scarcity Number of water rationing cases per year Urban Audit
C P 36%-‐38%
Yes, 3 years
Until a more meaningful indicator is available
UF12
Waste
Waste Intensity Annual amount of solid waste collected on urban territory per capita
Urban Audit
C P 53%-‐61%
Yes, 3 years
UF13
Recycling Share of solid waste recycled Urban Audit
C P 53%-‐61%
Yes, 3 years
UF14
Waste Treatment: Incineration
Share of solid waste incinerated Urban Audit
C P 53%-‐61%
Yes, 3 years
UF15
Waste Treatment: landfill
Share of solid waste landfilled Urban Audit
C P 53%-‐61%
Yes, 3 years
UF16
Land Soil sealing Increase in soil sealing on urban territory by type of converted land over the last five/ten years
GIS P Yes, 5 years
To be derived within IUME activities or from MOLAND
25 | P a g e
ID Area Name Unit Cities Barcelona Freiburg Lille Malmo Sofia
UF 1 Clim
ate & ene
rgy
CO2 intensity of production kg of CO2 year
Value: 0.09 9.3 (only total for this report all data available)
5.2 (only total for this report all data available)
Unavailable Years: 1990, 1995, 2006,
2007 UF 2 CO2 intensity of residential
activities Tonnes of CO2 emissions per capita per year
Value: 2.5 Years: 1990, 1995, 2006,
2007 UF 3 CO2 intensity of
transportation Tonnes of CO2 emissions per capita per year
Value: 1.5 Years: 1990, 1995, 2006,
2007 UF 4 CO2 intensity of energy use Tonnes of CO2 emissions per capita
per year Value: 9.4 Years: 1990, 1995, 2006,
2007 2004 2004
UF 5 Energy efficiency of production
Kilowatt hours per unit of GDP per year
Values: 0.21 6.7 (only total available MWH/cap)
Available Available Years: 2004, 2005, 2006,
2007, 2008
UF 6 Energy efficiency of transportation
Mega watt hours per capita per year Values: 2.8 Available Available Years: 2004, 2005, 2006,
2007, 2008
UF 7 Energy efficiency of residential usage
Mega watt hours per capita per year Values 3.0 Available Available Years 2004, 2005, 2006,
2007, 2008 2006
UF 8
Water
Territorial water extraction Percentage of water used on urban territory
Values Available 4 Unavailable Available Not clarified Years 1998, 2004, 2007
UF 9 Water use efficiency Cubic metres of water used per capita per year
Values 71.9 58.14 44.3 131.76 119.34 Years 1991, 1996, 2001,
2004 1991, 1996, 2001, 2004
2004 1996 2001, 2006, 2007, 2008
UF 10
Waste water treatment % dwellings connected to sewage system
Values 97.76 96.6 84.25 99.53 79.06 Years 2001, 2004 2001, 2004 2001 2001, 2004 2001
UF 11
Water scarcity Number of days (water rationing cases)
Values 0 0 -‐ 0 0 Years 2004 2001, 2004 -‐ 2001, 2004 2004
UF 12
Waste
Waste intensity Tonnes of waste per capita per year Values 0.58 0.46 -‐ 0.42 0.84 Years 2001, 2004 2007 -‐ 2004 2001
UF 13
Recycling Share (%) of solid waste recycled Values 24.9 45.0
1.8 -‐ 17.7
Years 2004 2007 1998 2004 2001
UF 14
Waste Treatment: Incineration
Share (%) of solid waste incinerated Values 18.2 32.3 98.2 81.4 49.7 Years 2004 2007 1998 2004 2001
UF 15
Waste Treatment: landfill Share (%) of solid waste landfilled Values 56.9 0 0 3.1 32.4 Years 2004 2007 1998 2004 2001
UF 16
Land Soil sealing Increase of soil sealing in m2 over the last 5/10 years
Values Unavailable Unavailable Unavailable Unavailable Unavailable Years
Table 3 Proposed urban flow indicators (Available data has been calculated, but not provided in time; Unavailable data request did not lead to positive response)
26 | P a g e
4.1.2 Urban Drivers Indicators on urban flows are at the core of the proposed indicator system. The first of three groups
driver category are aimed to provide relevant information on why we might observe changes in the physical metabolism over time or why we might see differences in the physical metabolism across cities. The latter is particularly interesting as it might lead to an identification of a set of generic factors determining (differences in) the physical metabolism of cities in Europe. We devise indicators on urban drivers across six thematic areas:
Population and households: This thematic area captures the developments in population and household size, population dynamics and household structure, which are important
related to population size, population growth, number of households and household size. There is considerable agreement that information on city size in terms of population is paramount for understanding urban systems. Cross-‐city studies have found interesting scaling relationships between different attributes of cities, notably resource consumption (Bettencourt et al. 2007). For example, infrastructures such as road networks usually scale sublinearly with city size, i.e. each additional citizen requires less than average additional infrastructure investment. However, total electricity consumption scales supralinearly with city size, i.e. additional dwellers consume more than the average. In this sense the size of a city measured in terms of its number of residents is an indispensable component for describing the urban system. Evidence further suggests that the demographic structure of a population can also determine the size and make-‐ (e.g. Haq et al. 2007). Even though we have opted not to include such an indicator so far, there is sufficient data to do so.
Lifestyles: We understand lifestyle broadly as they way in which residents of a city live and
metabolism (e.g. Baiocchi et al. 2010) and they must be expected to be of equal importance at the city level. We propose three simple indicators to capture key aspects of urban
can be expected to be closely related to the wealth of its citizens, employment opportunities etc.. Income provides some insights into the monetary resources people have available for consumption and the average occupancy per occupied dwelling shows provides how life is organized within the available physical infrastructure. In fact, we have shown elsewhere for the UK and Germany that the reduction in household size and dwelling occupancy are more important driver of CO2 emissions than population growth (Minx 2008; Baiocchi and Minx 2010). This finding is likely to hold for other aspects of the physical metabolism. Note that there are further possibilities to include indicators for social stratifications across groups of residents.
Local climatic conditionsmetabolism. Cold winters, for example, are one factor influencing heating requirements. The annual amount of rainfall influence available freshwater resources, irrigation requirements etc.. We include indicators on temperature and rainfall.
Prices: Price is one major factor influencing demand. We have included two price indicators
27 | P a g e
for a dwelling will influences who can ultimately afford to live in a city and how much room is affordable. Water price is one important factor determining how water is used in the city.
Transportation: How people move in a city between places is another important aspect
pollution and greenhouse gas emissions in a city. In some seminal contributions Newman and Kenworthy (1989; 1995; 1996) have, for example, identified a strong relationship between land-‐use and energy-‐use in urban systems. We propose five indicators in the area of transportation covering the overall level of transport demands, modal split, travel time, commuters into and out of the city as well as car ownership. Note that transport
Capacity for environmental regulation: Finally we propose to include an indicator on the
capacity of the city authority for environmental regulation;
For simplicity we propose to source the information almost exclusively from the urban audit in this indicator group. The only two exceptions where no data could be taken from the urban audit where an indicator for traffic volumes in cities and an indicator providing information on the capacity of city authorities for environmental regulation.
For the former the challenge is to find suitable data sources for a sufficiently large number of cities without having to rely on city cooperation. There are several smaller city level data sets on
Moreover, some data has been collected in European initiatives such as the European Common Indicators (Ambiente Italia Research Institute 2003).9 However, such databases are usually limited in terms of the number of cities included, can be commercial and are not easily comparable. The lack of comparable data at the city scale on transport demands in cities has been acknowledged by the European Commission and is currently being addressed in a comprehensive study on urban transportation (Rommers, 2010). At this moment in time we therefore suggest to continue the scoping of better transport data (represented in indicator UD12). Until better information is available for a large sample of cities, it seems best to stick with the mainly commuting transport information covered in the urban audit. An adequate indicator and data source for UD 17 still need to be found.
as shown in Table 5: data was usually found in the urban audit for multiple years for each city and data was also usually available for a common year across cities.
For the remaining twelve indicators data for all cities is only available in three instances, five times data was found for four of the five cities and in two cases for three cities.
This picture resembles well the situation for other urban audit cities. In the thematic area
between 68% and 71%. For the other thematic areas the ranges are more between 40% and 60%. In terms of the spatial delineation, for 11 out of the 16 variables data are provided at both larger urban 9 An evaluation of the state of the project still needs to be undertaken. It seems that the initiative is not on-‐going anymore. A website, which is supposed to provide data from the project, only contains a few samples.
28 | P a g e
zone and administrative level. The remaining five variables are only available for administrative boundaries.
29 | P a g e
ID Area Name Description Data Other data sources, comments
Source Spatial Unit
Availability Continuity Case
studies other
UD1
Popu
latio
n &
households
Population size Total number of resident living on city territory Urban Audit C,L 5/5 82%-‐97%
Yes, 3 years
UD2 Population growth Total population increase over the last 5 years Urban Audit C,L 5/5 Fill in Yes, 3 years UD3 Household Total number of households Urban Audit C,L 5/5 57%-‐
82% Yes, 3 years
UD4 Household size Average number of people per household Urban Audit C,L 5/5 57%-‐82%
Yes, 3 years
UD5
Lifestyle
GDP Gross domestic product at city level Urban Audit C,L 5/5 46%-‐81%
Yes, 3 years
UD6 Income Median disposable annual household income Urban Audit C,L 4/5 32%-‐40%
Yes, 3 years
UD7 Dwelling occupancy Average occupancy per occupied dwelling Urban Audit C,L 5/5 Yes, 3 years UD8
Local
clim
ate
Temperature Average temperature of the warmest and coldest month Urban Audit C 5/5 69%-‐71%
Yes, 3 years
UD9 Rainfall Average annual rainfall Urban Audit C 5/5 68%-‐70%
Yes, 3 years
UD10
Prices
House Prices Average house price Urban Audit 4/5 41%-‐47%
Yes, 3 years
UD11 Water Price Average water price Urban Audit C, L 4/5 42%-‐63%
Yes, 3 years
UD12
Transportatio
n
Traffic Volume Total number of vehicle kilometre per capita Cities C 1/5 Unknown Ideally expressed per km of road
UD13 Modal split Percentage of trips to work by mode Urban Audit C,L 3/5 Yes, 3 years UD14 Travel time Average travel time to work Urban Audit C,L 4/5 Yes, 3 years UD15 Commuter Net commuters into the city (Commuters into city minus
commuters out of city) Urban Audit C 5/5 Yes, 3 years
UD16 Car Ownership Number of private cars registered per capita Urban Audit C,L 4/5 Yes, 3 years
Table 4 -‐ Overview of proposed indicators for urban drivers (C=city level; L=larger urban zone)
30 | P a g e
ID Area Name Unit Cities Barcelona Freiburg Lille Malmo Sofia
UD1
Popu
latio
n & Hou
seho
lds
Population size Number of people in 1000 Value 1504 208 1091 260 1091
Years 1991, 1996, 2001, 2004
1991, 1996, 2001, 2004 1991, 2001, 2004 1991, 1996, 2001,
2004 1991, 1996, 2001
UD2 Population growth Percentage growth over the last five years
Value 0.98 1.26 0.13 0.94 -‐0.38 Years 1996, 2001, 2004 1996, 2001, 2004 2004 1996, 2001, 2004 1996, 2001
UD3 Households Number of households in 1000 Value 594 113 421 133 430
Years 1991, 1996, 2001 1991, 1996, 2001, 2004 1991, 2001, 2004 1991, 1996, 2001 1991, 2001
UD4 Household size Number of people per household
Value 2.53 1.85 2.59 1.95 2.54
Years 1991, 1996, 2001 1991, 1996, 2001, 2004 1991, 2001, 2004 1991, 1996, 2001 1991, 2001
UD5
Lifestyle
GDP Gross domestic product in
Value 40146 6847 51988 9915 4195
Years 1996, 2001, 2004 1991, 1996, 2001, 2004 2001 2004 2001
UD6 Income Median disposable annual household income i
Value 14200 19000 13183 21806 -‐
Years 1996, 2001, 2004 1991, 1996, 2001, 2004 2001 2001 -‐
UD7 Dwelling occupancy
Average occupancy per occupied building
Value 2.5 2.1 2.6 1.9 2.2
Years 1996, 2001 1991, 1996, 2001, 2004 1991, 2001, 2004 2004 2001
UD8
Local clim
ate Temperature Average temperature of the warmest and coldest month in degrees Celcius
Value 25.6 9.2
21.9 1.9
18.7 4.5
18.2 -‐1.7
22.4 -‐5.1
Years 1991, 1996, 2001, 2004 2001, 2004 2001, 2004 2004 2001
UD9 Rainfall Average annual rainfall in litres per square metres
Value 487 1125 862 697 519
Years 1991, 1996, 2001, 2004 2001, 2004 1996, 2001, 2004 2004 2001
UD10
Price levels House Prices Average house price in Euros
Value 2500 2700 1200 1468 -‐
Years 1991, 1996, 2001, 2004
1991, 1996, 2001, 2004 1996, 2001 1991, 1996, 2001,
2004 -‐
UD11 Water Price Average water price in euros per cubic metre
Value 1 1.7 3 0.7 -‐ Years 2001, 2004 2001, 2004 2004 2001, 2004 -‐
31 | P a g e
Continued
ID Area Name Unit Cities Barcelona Freiburg Lille Malmo Sofia
UD12 Transportatio
n Traffic Volume Total vehicle kilometres of
automobiles per capita Value -‐ 4913 -‐ -‐ -‐ Years -‐ 2005 -‐ -‐ -‐
UD13 Car travel Percentage of people commuting to work by car
Value 31.6 61.2 -‐ 51.0 -‐
Years 1991, 1996, 2001, 2004
1991, 1996, 2001, 2004 -‐ 2001, 2004 -‐
UD14 Travel time Length of trip in minutes Value 26.7 19.5 18.7 28 -‐
Years 2001, 2004 1991, 1996, 2001, 2004 1996 2001, 2004 -‐
UD15 Commuter Number of commuters per day Value 266246 35941 54651 33945 45810
Years 1991, 2001, 2004 1991, 1996, 2001, 2004 1991, 2001, 2004 1996, 2004 2001
UD16 Car Ownership Average number of cars per capita
Value -‐ 0.34 0.38 0.37 0.52
Years -‐ 1991, 1996, 2001, 2004 1991, 2001, 2004 1991, 1996, 2001,
2004 2001
Table 5 -‐ Urban Driver Indicators for five test cities
32 | P a g e
4.1.3 Urban Patterns
indicators tries to capture important aspects of land-‐use and the built environment in cities, which might be key for understanding infras .
One challenge associated with the proposed indicator system is that urban flow indicators are only available at the city level (administrative boundaries). In terms of urban patterns, city-‐level data usually does not provide sufficient information about how the city might be shaped within its administrative boundaries. Simple indicators such as population density, for example, can therefore provide a very biased picture of land use if they are calculated at a high aggregation level. We therefore take a two-‐tiered strategy in sourcing information for this group of indicators. First, we use some variables from the urban audit to describe general features of the city at the administrative level. Second, we recommend the construction of specific indicators from more detailed GIS data sets to characterise specific urban form and land-‐use aspects within the administrative area of a city. For the latter a variety of indicators have been suggested in the literature. An excellent review is included in the analysis of European cities by Schwarz (2010).
For indicators of urban form we follow Huang (2007) in this report, who distinguishes four types of spatial metrics relevant for characterizing cities related to complexity, centrality, compactness and porosity as shown below.
Measures of complexity try to capture the regularity of the patch shape, i.e. how the borders of the sealed urban patch within the administrative boundaries of the city are shaped. Measures of centrality indicate distance of urban development to some defined centre such as the central business district. Compactness measures try to capture patch shape and fragmentation of the overall urban landscape. Finally, porosity indicates the ratio of open space compared to the total sealed urban area.
Box 3 -‐ Urban form indicators as proposed by Huang (2007)
33 | P a g e
in
City size: The size of the urban territory in terms of square kilometres is a basic variable for understanding the spatial extent of the area under consideration.
Land cover and land use: This thematic area gives insights into how the city territory is used for different purposes. This is not only important for aspects of urban ecosystem service provisioning, but can also be of direct relevance for getting a better grasp on the physical make-‐up of the administrative area, which can influence its physical metabolism. Note that
is the opposite of the porosity measure introduced above (see Box 3).
Transportation network: The importance of transportation in the context of size and shape
provides important monetary and non-‐monetary incentives for modal choice of city residents. We include two indicators measuring the share of the different transport infrastructures by mode on urban land-‐take as well as the length of the public transport network.
Urban form: We have already motivated the importance of urban form indicators for understanding the urban metabolism at the beginning of this Section. In this thematic area we propose three indicators: compactness, centrality and population density. How the compactness and centrality measures have been calculated is indicated in Box 4 below. Further note that different population density metrics are available. Currently the measure refers to the total urban territory. It could be more appropriate to relate population only to the sealed urban area.
Buildings: Finally the number of dwellings (and changes in) gives another indication of how the urban built-‐up environment develops.
We use two types of data sources. For most of the indicators urban audit was used as a data source. As for previous indicator groups data availability varies. For three of the urban audit indicators data is available for all cities for a common year (city size, population density, building stock) and for one indicator for all cities but different years (length of transport network). For the land use and land cover variables the data situation is more difficult. Two indicators (built-‐up land, open space) are only available for four and one indicator (transport land) for three of the five cities.
This pattern also captures quite well data availability for the other urban audit cities. The variables -‐85% of the
urban audit cities, while the land-‐use and land cover variables are available for 35% to 50% only.
lity to construct these variables from databases such as CORINE or Urban Atlas.
The remaining indicators for soil sealing, compactness and centrality were constructed from CORINE by Nina Schwarz (2010), which we would like to acknowledge gratefully for providing this data to us. These measures could be constructed from available databases from Urban Atlas for all larger urban zones in Europe with a resident population larger than 100000 inhabitants. Also in terms of the
34 | P a g e
spatial delineation of urban areas, there is a large degree of freedom for land-‐use and land cover related variables whether to follow an administrative, functional or morphological approach.
Following Huang (2007) and Schwarz (2010) centrality measures the average distance of the dispersed sealed urban patches on the city territory to the city centre, which is defined as the centroid of the largest patch. Let Di represent the distance of the centroid of patch i to the centroid of the largest urban patch, N the total number of patches, R the radius of a circle with the area of s and s the total area of all patches. We can then calculate a centrality index by
Compactness measures patch shape and fragmentation of the overall urban landscape. Let si and pi represent the area and perimeter of patch i, Pi the perimeter of a circle with the area of si and N is the total number of patches. We can then calculate a compactness index by
(Huang et al. 2007). Box 4 -‐ Definitions of centrality and compactness
35 | P a g e
ID Area Name Description Data Other data sources, comments Source Spatial
Unit Availability Continuity
Case studies
other
UP1 Size
City size Spatial extent of city according to cadastral register (in km2)
Urban Audit C,L 5/5 66%-‐
68% Yes, 3 years Various
UP2
Land
cover and
use
Sealed Land Percentage of sealed urban area of city size (%) CORINE C,L 4/5 Yes, 5 years Urban atlas; inverse of sealed land is porosity indicator
UP3 Built-‐up land Share of land used for residential and commercial purposes (%)
Urban Audit C,L 4/5 35%-‐
45% Yes, 3 years Urban atlas, CORINE
UP4 Open spaces Share of green spaces area, water and wetland Urban Audit C,L 4/5 38%-‐
52% Yes, 3 years Urban atlas, CORINE
UP5
Transpor
t
Transport Land Share of land used for transport (road, rail, ports) (%) Urban Audit C,L 2/5 Yes, 3 years Urban atlas
UP6 Transport network length
Length of public transport network per inhabitant (km per capita)
Urban Audit C 5/5 Yes, 3 years Urban atlas
UP7
Urban Form Compactness Index The compactness index measures the individual patch
shape and fragmentation of the landscape. CORINE C 4/5 Yes, 5 years
UP8 Centrality Index The centrality index indicates the average distance of sealed urban patches with respect to the largest patch CORINE C 4/5 Yes, 5 years
UP9 Population density Density of population in relation to city size Urban Audit C 5/5 66%-‐
68% Yes, 3 years There are alternative density indicators
UP10
Build
ings
Building Stock Total number of dwellings (houses, apartments) Urban Audit C 5/5 66%-‐
85% Yes, 3 years
Table 6 -‐ Overview of proposed indicators for urban patterns (C=city level; L=larger urban zone)
36 | P a g e
ID Area
Name Unit Cities Barcelona Freiburg Lille Malmo Sofia
UP1 Size City Size Spatial extent of city according to cadastral register in km2
Values 99 153 606 154 451 Years 1991, 1996,
2001, 2004 1991, 1996, 2001, 2004
2001, 2004 1991, 1996, 2001, 2004
2001
UP2
Land
cover and
use
Sealed Land Percentage of sealed urban area of city size Values 0.79 0.24 0.41 0.41 -‐ Years 2000 2000 2000 2000 -‐
UP3 Built-‐up land Percentage of land used for residential and commercial purposes
Values Res: 33.1 Com: -‐
Res: 9.4 Com: 2.6
Res: 28.6 Com: 11.2
-‐ Res: 23.7 Com: -‐
Years 1991, 1996, 2001
2001, 2004 1996, 2001 -‐ 2001
UP4 Open spaces Share of green spaces area, water and wetland areas
Values -‐ Gs: 70.6 Ww: 2.4
Gs: 2.2 Ww: 1.3
Ww: 0 Ag: 37.9
Gs: 30.6 Ww: 10.6
Years -‐ 1996, 2001 2001 2001 UP5
Transpor
t
Transport Land Percentage of land used for transport (road, rail, ports)
Values -‐ 10.0 1.0 -‐ -‐ Years -‐ 2001, 2004 2001 -‐ -‐
UP6 Length of Transport Network
Length of public transport network in kilometres per capita
Values 0.6 1.5 1.1 0.7 0.7 Years 1996, 2001 2001, 2004 2001, 2004 2004 2001
UP7
Urban fo
rm
Compactness Index Index (higher for more compact developments) Values 0.09 (0.49)
0.04 (0.31)
0.01 (0.12)
0.04 (0.19)
-‐
Years 2000 2000 2000 2000 -‐ UP8 Centrality Index Index (the larger the less centralized a city) Values 2.5 2.28 67.58 1.33 -‐
Years 2000 2000 2000 2000 -‐ UP9 Population density Number of residents per square kilometre Values 15232 1361 1787 1673 2532
Years 1991, 1996, 2001, 2004
1991, 1996, 2001, 2004
2001 1991, 1996, 2001, 2004
2001
UP10 Buildings
Building Stock Total number of dwellings (houses, apartments)
Values 757928 97872 441707 137647 468699 Years 1991, 2001,
2004 1991, 1996, 2001, 2004
2001, 2004
2001, 2004
2001
Table 7 -‐ Urban Pattern Indicators for five test cities
37 | P a g e
4.1.4 Urban Quality Improvements in the quality of life in cities henceforth referred to as urban quality -‐ should be the aim of all urban policies. Monitoring how changes in the physical metabolism of cities might be related to aspects of urban quality are therefore key for monitoring the success of and scoping the need for urban policies. How to operationalise and monitor quality of life is rooted in a longstanding academic debate. While some authors have tried to devise a single, overall welfare measure, there has been an increasing agreement that the comprehensiveness of welfare, well-‐being and quality of life concepts requires multiple indicators drawn together in an indicator system (see Keuning 1994; Keuning et al. 1999; Stahmer 2000).
We have drawn-‐up a system of fourteen indicators in four thematic areas:
Air Pollution: Even though local air pollution is an integral component of the urban metabolism, it is also a major determinant of quality of life in a city due to its detrimental effects on human health. As we perceive the significance of urban air pollution for decision-‐making to be related to these health effects, we have opted to include these indicators in
ndicators in this thematic area covering the short and long-‐term aspects of ozone, nitrous dioxide and particulate matter.
Noise: Noise has a significance impact on quality of life and must be considered as a health (defined as physical and mental well-‐being and the absence of disease) threat according to the World Health Organization (Suter 1991). Because many of the noise problems in cities are related to urban transportation activities, it potentially offers large potential for improving urban quality whilst reducing the physical metabolism. We include two indicators related to the noise exposure of city residents during day and night times.
Infrastructure, green space and accessibility: The quality of the available infrastructure, the access to the city as well as its green spaces are other important aspects of urban quality. However, the relationship to the physical metabolism of a city is not always straightforward. While retrofitting houses in a bad condition can save a lot of energy, the improvement of the accessibility to a city might trigger additional transport demands. We include five indicators in this thematic area related to the quality of the water and housing infrastructure, access to green spaces and availability of areas for recreation and leisure as well as accessibility.
Social aspects/unemployment: There are other socio-‐economic factors influencing the well-‐being of people. Studies have shown, for example, that unemployment and bad health are
reported happiness than income (Frey and Stutzer 2002). This is currently the only socio-‐economic variable included here. However, in the future it might be desirable to extend this indicator component.
There are a few areas where this indicator system might need completion. First, we have not included health related indicators so far. Second, there is the need to consider the inclusion of data reflecting the residents perception of quality of life in the city.
In terms of data sources most variables have been taken from the urban audit. Data availability for the other variables is mostly good as shown in Table 8 and Table 9. All local air pollution indicators are available for all five testbed cities for a common year. Indicators in the thematic area
38 | P a g e
Unemployment data is available for all cities, but different years. The major bottleneck in the urban audit is noise data. However, in this area information can be sourced from the NOISE database (http://www.eea.europa.eu/themes/noise/dm), where already city level data has been constructed. In the remaining Sections we will discuss how the data can be summarised and analysed to address some of the policy questions raised in this tender.
39 | P a g e
ID Area
Name Description Data Other data sources, comments Source Spat
ial Unit
Availability Continuity Case studies
other
UQ1
Air P
ollutio
n
O3 short Summer Smog: Number of days ozone (O3) concentrations exceed 120 microgram/m3
Urban Audit C 5/5 ~55% Yes, 3 years
UQ2 NO2 short Number of hours per year that nitrogen dioxide NO2 concentrations exceed 200 microgram/m3
Urban Audit C 5/5 39%-‐56% Yes, 3 years
UQ3 PM10 short Number of days particulate matter PM10concentrations exceed 50 microgram/m3
Urban Audit C 5/5 50%-‐61% Yes, 3 years
UQ4 O3 long Accumulated ozone concentration in excess 70 microgram/m3 Urban Audit C 5/5 40%-‐54% Yes, 3 years
UQ5 NO2 long Annual average concentration of NO2 Urban Audit C 5/5 42%-‐59% Yes, 3 years UQ6 PM10 long Annual average concentration of PM10 Urban Audit C 5/5 38%-‐57% Yes, 3 years
UQ7
Noise Noise day Proportion of residents exposed to traffic
noise during the day Urban Audit C 1/5 2%-‐6% Yes, 3 years NOISE provides alternative data
UQ8 Noise night Proportion of residents exposed to traffic noise at night Urban Audit C 1/5 2%-‐6% Yes, 3 years NOISE provides alternative data
UQ9
Infrastructure, green sp
ace
& accessib
ility
Water quality
Proportion of dwellings connected to potable drinking water system Urban Audit C 4/5 43%-‐54% Yes, 3 years
UQ10 Housing quality Average area of living accommodation Urban Audit C,L 5/5 38%-‐60% Yes, 3 years Indicator could be misleading
UQ11 Green space access
Green space to which the public has access Urban Audit C,L 4/5 38%-‐44% Yes, 3 years
UQ12 Recreational land
Proportion of land area in recreational, sports and leisure use Urban Audit C, L 5/5 ~36% Yes, 3 years
UQ13 Accessibility Multimodal accessibility Urban Audit C,L 5/5 Yes, 3 years
UQ14 Social
Unemployment rate
Residents unemployed as a share of all economically active residents Urban Audit C,L 5/5 49%-‐75% Yes, 3 years
Table 8 -‐ Overview of proposed indicators for urban quality (C=city level; L=larger urban zone)
40 | P a g e
ID Area Name Unit Cities Barcelona Freiburg Lille Malmo Sofia
UQ1 Air P
ollutio
n O3 short Number of days ozone (O3) concentrations exceed 120
microgram/m3
Value 3 26 13 0 2
Years 2004 1991, 1996, 2001, 2004 2001, 2004 1996, 2001,
2004 2001, 2004
UQ2 NO2 short Number of hours per year that nitrogen dioxide NO2 concentrations exceed 200 microgram/m3
Value 0 0 1 0 0
Years 2004 1996, 2001, 2004 2001, 2004 1996, 2001,
2004 2004
UQ3 PM10 short Number of days per year particulate matter PM10concentrations exceed 50 microgram/m3
Value 66 9 3 1 112
Years 2004 1996, 2001, 2004 2001, 2004 1996, 2001,
2004 2004
UQ4 O3 long Accumulated ozone concentration in excess 70 microgram/m3 in microgram per square meter
Value 2065 4834 3092 2264 1920
Years 2004 1996, 2001, 2004 2001, 2004 1996, 2001,
2004 2001, 2004
UQ5 NO2 long Annual average concentration of NO2 in microgram per square meter
Value 47.5 21 30.9 19,5 29.2
Years 2004 1996, 2001, 2004 2001, 2004 1996, 2001,
2004 2004
UQ6 PM10 long Annual average concentration of PM10 in microgram per square meter
Value 38.8 18.5 23.1 15.9 50
Years 2004 1996, 2001, 2004 2001, 2004 1996, 2001,
2004 2001, 2004
UQ7
Noise Noise day Proportion of residents exposed to traffic noise during the
day Value -‐ -‐ -‐ 28.6 -‐ Years -‐ -‐ -‐ 2004 -‐
UQ8 Noise night Proportion of residents exposed to traffic noise during the night
Value -‐ -‐ -‐ 35.1 -‐ Years -‐ -‐ -‐ 2004 -‐
41 | P a g e
Continued
ID Area Name Unit Cities Barcelona Freiburg Lille Malmo Sofia
UQ9
Infrastructure, green sp
ace &
accessibility
Water quality Proportion of dwellings connected to potable drinking water system
Value 97.5 100 -‐ 100 -‐ Years 2001, 2004 2001, 2004 -‐ 2001, 2004 -‐
UQ10 Housing quality Average area of living accommodation in square meter per person
Value 34 36.9 36.1 42 14.6
Years 2001, 2004 1991, 1996, 2001, 2004 2001 2001 1991, 2001
UQ11 Green space access
Green space to which the public has access in square metre per capita
Value 4.1 304 -‐ 93.5 169.2
Years 1991, 1996, 2001, 2004 2001, 2004 -‐ 2004 2001
UQ12 Recreational land Proportion of land area in recreational, sports and leisure use
Value 2.6 3 1.2 4.5 1 Years 1991, 2004 2001, 2004 1997, 2001 2004 2001
UQ13 Accessibility Index, where 100 represents EU27 average Value 127 124 120 126 99 Years 2004 2004 2004 2004 2004
UQ14
Social
Unemployment rate
Residents unemployed as a share of all economically active residents
Value 11.95 8.49 14.42 10.55 4.32
Years 1991, 1996, 2001, 2004
1991, 1996, 2001, 2004 2001 2001 2001
Table 9 -‐ Urban Quality Indicators for five test cities
42 | P a g e
4.1.5 Headline indicator set We propose a headline indicator set as a descriptive summary of the information compiled. The headline indicator set is summarised in Table 10. It consists of 15 indicators across the four proposed dimensions (urban flows, urban patterns, urban drivers and urban quality) of the proposed metabolism concept.
No ID Name Dimension H1 Per capita CO2 emissions from energy consumption Urban Flows H2 Energy efficiency of transport Urban Flows H3 Efficiency of residential energy use Urban Flows H4 Efficiency of urban water use Urban Flows H5 Waste intensity Urban Flows H6 Recycling Urban Flows H7 Urban land take Urban Flows H8 Green space access Urban Quality H9 NO2 concentrations Urban Quality H10 PM10 concentrations Urban Quality H11 Unemployment rate Urban Quality H12 Land use efficiency Urban Patterns H13 Public transport network length Urban Patterns H14 Registered cars Urban Drivers H15 GDP per capita Urban Drivers
Table 10 -‐ The proposed headline indicator set
Radar charts, as shown in Figure 12 for the cities of Malmo, Freiburg and Barcelona, provide an effective way of summarising information. We have normalised the data to the sample average. This means that a value larger than 1 indicates that an attribute is more developed than for the sample average and a value smaller than 1 that it is less developed. Barcelona, for example, is much denser than in Malmo and Freiburg and is lower in CO2 emissions per capita. Freiburg is denser than Malmo, but higher in per capita CO2 emissions. This might be partially explained by the fact that Freiburg has a high share of car use in its modal split (See discussion on the issue of spatial aggregation level for further information). The citizens of Malmo and Freiburg enjoy much lower local pollution levels than people in Barcelona and have much better access to public green spaces, which are important aspects of urban quality of life. Note that Figure 12 mainly serves for demonstrative purposes. Given the current data availabilities it was not possible to construct exactly the indicators outlined in Table 10 particularly the energy and CO2 indicators. However, it suffices to depict the type of information that can be obtained.
43 | P a g e
Figure 12 -‐ Headline indicators for three test cities. A value bigger than 1 means that the attribute is more developed than in the average European city. A value smaller than 1 means that an attribute is less developed than for the average
city in the sample. A zero value indicates data unavailability.
4.1.6 Applications With such a headline indicator set, we can obtain a general overall picture of the cities under consideration. However, even if we have information on a larger sample of cities, it is unlikely that we are able to answer some of the questions, which define the scope of this service contract. This is partially due to limited data availability (time series, spatial aggregation, consumption based indicators) and partially due to the fact that we are interested in the identification of general trends associated with urbanisation in Europe. With regard to the latter the descriptive indicator statistics are often insufficient, because we are often interested in the relationship between and its metabolism or the identification of fundamental relationships that ex. For example, how is
2 emissions (or gasoline consumption)? What are the primary drivers and structural features of cities, which determine their energy consumption? We therefore propose two analytical extensions:
First, it is intriguing to look beyond a specific city and analyse how changes in social organisation and dynamics resulting from urbanisation will impact the interactions between nature and society. In particular, authors have proposed that there are fundamental scaling relationships between the size of cities, economic development and knowledge creation and provided evidence indicating that these relationships are quantitatively consistent across nations and time (Bettencourt et al. 2007). We propose to analyse the existence of such a set of general properties for European cities with a particular focus on resource consumption: how the size of cities relates to infrastructure investment
44 | P a g e
and electricity or water consumption might provide important insights into the drivers behind these relationships and opportunities for targeted policy interventions. In order to summarise some metabolic implications from urbanisation processes, we propose to introduce an aggregate indicator table for urban systems that compiles diverse scaling relationships between population size and key environmental, infrastructural and social factors as shown in Fehler! Verweisquelle konnte nicht gefunden werden.. In the future this could also provide the opportunity to better understand the social contribution cities make to knowledge and economic development.
No Name Description A Road length Total length of the urban road system B Length public
transport network Total length of the public transport system
C Car traffic volume/ fuel consumption
Total automotive fuel consumption on urban territory
D Buildings Number of buildings in the urban area E GDP Gross domestic product of the city F Water extraction Water extracted on urban territory G Water consumption Water consumed on urban territory H Electricity
consumption Electricity consumed on urban territory
I Air pollution Exposure to air pollution on an urban territory J Waste collection Total amount of waste collected.
Table 11 -‐ Example for scaling relationships that could be summarised in a headline indicator set for urbanisation in Europe
Second, in order to gain a deeper understanding of the urban metabolism in Europe we further propose the application of multivariate regression techniques to analyse how physical attributes of the metabolism of European cities are related to a well selected set of urban drivers, patterns and quality of life aspects. By doing so, we can identify key influences on the urban metabolism. The specification of these regression models will depend on the research question under consideration
IUME initiative.10 In fact, by carrying out such an analysis the three pillars (data, qyestions, systems) of IUME are integrated. It is not possible to carry out this research under the current service contract due to the absence of a sufficiently large set of data (only 5 test cities considered here) and limited project resources.
4.1.7 Towards an urban metabolism database We previously mentioned that the proposed indicators in the urban flow dimension are taken from an urban metabolism database. This database does not exist yet and should be constructed drawing from a variety of existing data sources. This database will arranged around five themes:
Climate & energy; Water; Waste; Land; Local and regional air pollution.
10 see, http://iume.ew.eea.europa.eu/concept/questions
45 | P a g e
the five themes will need to be structured as exemplified in Box 2 for the water cycle. This will provide a structure for systematically collecting and storing data related to a particular aspect of the metabolism.
We do not perceive this database development to be particularly resource intensive. Hence, it is our recommendation to derive the indicators from existing databases such as WISE, CORINE, Urban Atlas, Soil Sealing, EPER etc.. In many cases indicators one will be able to readily transfer the indicators into the urban metabolism database, in other cases some manipulations might be required (e.g. transformations, imputations, downscaling etc.; see Figure 13). Once a suitable database structure is in place, database development activities undertaken within the EEA or IUME can automatically feed into it. It is probably most challenging to ensure that data is submitted. This is one of the reasons why we propose in the last Section of this report the establishment of a working
developments in the thematic areas at the EEA or related institutions. In the future, additional standard indicators might be derived from this database (e.g. indicator set for city governments) and extensions from thematic towards functional descriptions (e.g. housing; mobility) of the urban metabolism might be considered.
Figure 13 -‐ The Urban metabolism database
46 | P a g e
4.2 Approach 2: Small area estimates for carbon footprints and energy consumption
There are two main restrictions associated with the indicator system outlined above. First, there are severe data restrictions when it comes to the calculation of consumption-‐based indicators, which capture the system-‐wide environmental pressures generated throughout the global supply chains of goods and services consumed in cities. Second, the proposed indicator system works at the administrative city level only. This poses a variety of challenges. For example, for some types of analyses administrative delineations of cities might not suitable for analysis and functional or morphological delineations might be the more appropriate choice. Fons et al. (2008) argue that environmental issues are often best analysed based on morphological delineations. Moreover, several of the issues the European Environmental Agency would like to shed light on with an urban metabolism concept, require more detailed and a wider set of data. Fundamental questions such as whether cities actually saves or triggers additional resources/pollution compared to non urban areas require information about other types of human settlements and rural lifestyles. Similarly, for analysing sprawl or urbanisation processes in depth, more detailed information of the sprawling areas could be of great help.
Hence, this Section discusses opportunities and challenges associated with using more comprehensive data for analysing the urban metabolism. We use comprehensive consumption based estimates of CO2 emissions in the UK. The data set covers the whole country and was obtained using a downscaling methodology. The merits of such a downscaling approach will therefore be discussed as well.
4.2.1 General introduction The challenge of calculating consumption-‐based pollution and resource use accounts is the requirement to combine information on international trade and global production activities, on the one hand, with information on local consumption activities on the other hand. In the literature on
surge of interest in establishing full consumption based accounts (Ramaswami et al. 2008; Kennedy et al. 2009; Hillman and Ramaswami 2010; Kennedy et al. 2010; Minx et al. 2010; Parshall et al. 2010). However, many of the proposed bottom-‐up methods are very work intensive and can usually only be established for individual cities. This is similar to existing experiences in the urban metabolism literature including Ecological Footprints.
Another route that can be taken is to try and downscale information carbon footprint estimates to smaller spatial scales. The question is whether such downscaling exercises can lead to meaningful results and what additional insights the results might provide for addressing policy issues raised in the context of this tender. The national consumption based emission CO2 and GHG estimates marking the starting point for this downscaling exercise have gone through peer review on multiple occasions (Baiocchi and Minx 2010; Wiedmann et al. 2010), while this is only be partially the case for the downscaling methodology itself (see Minx et al. 2009).
Global production activities in the model are represented in a global multi-‐regional input-‐output model (Baiocchi and Minx 2010; Wiedmann et al. 2010). Such models depict flows of goods and services between economic sectors, regions and final demand entities in monetary units. Adding a vector of emission intensities and assuming that each unit of sector output generates the same amount of CO2/GHG per unit of monetary output, we can assess the direct and indirect emission
47 | P a g e
requirements of a given final demand (or the carbon footprint of final consumption activities) at the national level.
To calculate carbon footprints at smaller spatial scales in such a model environment we need local final consumer expenditure data. The basic challenge associated with the construction of local expenditure matrices is insufficient sample size of national consumer expenditure surveys to obtain estimates at sufficiently small spatial scales even when data is pooled from multiple surveys. Geodemographic data can be used to downscale information.
Geodemographics is a discipline, which has been concerned with spatial downscaling of socio-‐economic (and other) information for a long time. It might be bes
knowledge about where people live reveals information about them (Harris et al. 2005). From an extensive survey usually a census a lifestyle classification is built in a bottom-‐up clustering procedure using a wide range of variables such as area characteristics, type of housing, income level or ethnicity etc.. Lifestyle types with similar characteristics are grouped together. In the end each street in the country is assigned to a dominant lifestyle type. Here we use the commercial MOSAIC UK classification provided by Experian Ltd, but there are other commercial and academic systems available.
Given the knowledge where lifestyle types tend to live across the country local consumption expenditure matrices can be imputed once the national consumer expenditure survey has been coded according to this lifestyle classification. However, such a procedure then assumes that there is no variability within lifestyle types regardless where people live. We try to overcome this assumption by updating our initial estimates at various spatial scales with the best information available. To make these updates as good as possible we build data hierarchies, where information at higher spatial level overrides lower level information (scaling procedures) and physical data overrides monetary data (updating procedures). Based on this methodology we are able to devise comprehensive consumption based CO2/GHG emission estimates for all local authorities (354) and middle layer super output areas (7194) in England.
4.2.1.1 Results In Figure 14 we show CO2 emission accounts of five cities in the UK from a production and a consumption based perspective: London, Manchester, Brighton, Milton Keynes and Hartlepool. The first important thing to note is that for all cities apart from Hartlepool the direct and indirect emissions required in the global production of goods and services finally consumed on their territory are larger than the CO2 emissions released. Part of the reason is that the UK as a whole is a net importer of CO2 emissions from the rest of the world (Baiocchi and Minx 2010; Baiocchi et al. 2010; Wiedmann et al. 2010), i.e. consumer emissions are higher than producer emissions on average. Another explanation is that cities often rely on their Hinterland in the production of goods and services. Many of the coal power plants in the UK, for example, are located in rural areas even though a considerable share of the electricity is consumed in cities. The local authority of North Lincolnshire where the largest coal fired power plant of the UK is located has producer emissions of 69 t CO2 per capita.
This brings us to the second point that consumer emission estimates show much less fluctuation. This means that consumption patterns are much evenly distributed across space than CO2 emission
48 | P a g e
sources. Producer and consumer emissions are essentially uncorrelated and applying one or the other depends on the policy question under consideration. Many questions in the debate on climate change in the context of cities are aimed at understanding a) whether and under which conditions urban life might provide CO2 benefits; b) how potential CO2 benefits associated with urban life can be reaped through intelligent urban planning. Such questions ask are directed to all aspects of life in a city and global, system wide CO2 emission releases and require consumption based accounting. Similarly, in the context of benchmarking the climate change performance of cities, authors have argued that the unequal distribution of CO2 emission sources makes the application of a producer emission accounts meaningless and that the establishment of consumer emission accounts is of upmost importance (Kennedy et al. 2009; Hillman and Ramaswami 2010).
Figure 14 -‐ CO2 emission of cities in the UK from a production and consumption perspective (Data from DECC and SEI)
Even if we manage to introduce some consumption based indicators into our indicator system introduced in the previous Section, there are some limitations associated with the fact that the system focuses on cities only. However, some questions associated with urbanisation processes, sprawl and the global environmental impacts of cities require comparisons with life in non-‐urbanised areas and therefore spatially more comprehensive accounts comprising also the rural parts of the country.
49 | P a g e
Figure 15 -‐ Per capita carbon footprint by degree of ruralness. A value of 1 represents highly urbanised areas, while a value of 6 represent highly rural areas.
One question, for example, is how different levels of urbanisation are related to direct and indirect greenhouse gas emissions associated with cities. In Figure 15 we show aggregated estimates of the carbon footprint from household consumption activities in rural and urban areas. We have derived these figures from consumption based CO2 emission estimates at local authority level. The data indicates that household consumption activities in rural areas trigger more CO2 emissions directly and indirectly around the globe than urban household consumption, even though the differences remain relatively small (within 10%). Looking at contribution of emission components in Table 12, it seems that the largest differences across urban and rural household are related to transportation and housing. In particular, rural households tend to have a higher footprint in these consumption categories. In general, Figure 15 further suggests that the global CO2 emissions from household consumption do not slowly increase with the degree of ruralness. Instead, there seem to be two clusters: CO2 emissions from -‐3) show very similar levels and patterns and are lower than CO2 households (1-‐6), which are also similar in levels and compositions.
Degree of ruralness Consumption area
Unit 1 2 3 4 5 6
Food & Drink Tonnes of CO2 per capita 1.1 1.0 1.0 1.1 1.1 1.1 Housing Tonnes of CO2 per capita 3.2 3.1 3.1 3.4 3.4 3.3 Transport Tonnes of CO2 per capita 2.5 2.8 2.8 3.0 3.0 3.1 Other Tonnes of CO2 per capita 2.3 2.3 2.3 2.4 2.4 2.4 Total Tonnes of CO2 per capita 9.1 9.2 9.2 9.9 9.9 10.0 Table 12 -‐ Carbon footprint by degree of ruralness. A value of 1 represents highly urbanised areas, while a value of 6
represents highly rural areas
The next question is whether this divide between consumer emissions in urban and rural areas is also reflected in a more gradual picture. The literature on cities and energy consumption/ CO2 emissions, for example, has been interested in the relationship with population densities (Newman and Kenworthy 1989; Kenworthy and Laube 1996; Newman and Kenworthy 1996). Figure 16 reveals
50 | P a g e
that there is no conclusive evidence at local authority level for this relationship. Even when moving to middle layer super output areas dividing England into more than 7000 spatial entities this relationship does not become more pronounced (graph not included for matters of space). The question therefore becomes whether this relationship simply does not exist, whether there are more subtle trends currently not captured or whether the quality of the imputed consumption based estimates are not of sufficient quality.
Figure 16 Relationship between consumption based CO2 emissions and population density at local authority level
It is another appeal of the data that it does not only provide spatial detail, but also detail across consumption categories as shown in Table 12. Overall, 44 categories of consumption can be distinguished. Even if we do not see a relationship between density and CO2 emissions at the aggregate level, there could be some emission components that do. The most obvious candidate is the area of transportation. Even though Newman and Kenworthy (1989; 1996) have shown a strong negative relationship between gasoline consumption and density for a global set of cities, European data has not been conclusive so far.
In Figure 17 we find a similar relationship in our data when we focus on direct and indirect CO2 emissions associated with private transportation: the denser the area the smaller the CO2 emissions associated with personal travel. Overall, about 60 percent of the variation in the data can be explained by this simple model. Interestingly, once we include all transport activities into our consideration as shown in Figure 18 we largely lose this relationship again (i.e. much less of the variation in the data can be explained). Moreover, instead of a downward sloping we obtain the best fit for a U-‐shaped relationship between population density and per capita consumer CO2 emissions.
51 | P a g e
Figure 17 Relationship between direct and indirect CO2 emissions from private transportation (excluding the purchase of motor vehicles)
Figure 18 Relationship between direct and indirect CO2 emissions from all personal transport activities including the purchase of motor vehicles and all other transport services
However, we are not the first who find that the inclusion of other transport aspects counteract the benefits reaped from denser urban developments. Holden and Norland (2005), for example, find that the carbon savings associated with everyday travel in more densely populated urban areas are offset by the fact that the population tends to have the highest leisure time travel undertaken by plane.
Figure 19 -‐ Relationship between average weekly household income and per capita CO2 emissions from consumption across local authorities in England
This leaves the question what drives the direct and indirect CO2 emissions across all consumption areas. What we typically find for studies of different lifestyle groups in a country is that income is a key driver of direct and indirect CO2/GHG emissions (Baiocchi et al. 2010). Figure 19 shows that this relationship also seems to hold for average weekly income in local authorities: the more people earn
52 | P a g e
on average in a local authority the higher its global climate change impacts. This result has been suggested before in the literature. However, given our geodemographic approach to downscale information there iaspect therefore of our attempt to evaluate the usefulness of the consumption based CO2 emission estimates provided, is to validate at least some of the findings with independent data.
However, before we do this let us briefly increase the spatial resolution of our estimates and show how this might increase the number of applications. Figure 20 to Figure 23 show the direct and indirect greenhouse gas emissions associated with consumption in London, Manchester, Hartlepool and Brighton at middle layer super output area (MLSOA) level.
Figure 20 -‐ The direct and indirect greenhouse gas
emissions associated with consumption in London at middle layer super output area
Figure 21 -‐ The direct and indirect greenhouse gas
emissions associated with consumption in Manchester at middle layer super output area
Figure 22 The direct and indirect greenhouse gas emissions associated with consumption in Hartlepool at middle layer
super output area
Figure 23 -‐ The direct and indirect greenhouse gas
emissions associated with consumption in Brighton at middle layer super output area
Moving to such a smaller geography has a variety of advantages for studying the environmental consequences associated with life in cities. First, at larger administrative levels averaging effects are prominent, which might hide important differences within these areas. For example, population density in London at local authority level varies between 20 and 130 people per square kilometre. Moving to MLSOA this range is between 0 and 19280. Second, more spatial detail offers opportunities to move between different urban delineations depending on the analytical requirements. Third, greater spatial detail might also offer the opportunity to focus on key areas of urbanisation and sprawl processes. Fourth, only such higher spatial resolutions of emission
53 | P a g e
estimates might enable us to identify which components of the social and physical environment determine metabolic flows within an urban system.
Again it is not the ambition here to provide a comprehensive analysis of the emission structure of our five English test cities. However, the Figures clearly show the added analytical options. Returning to our discussion on the importance of population densities for determining the direct and indirect GHG emissions from consumption, the pictures for Manchester (Figure 21) and Hartlepool (Figure 22) seem to suggest from a visual inspections that emissions at the urban fringes are higher than at the core. Lower densities could be one intuitive explanation for this.
Indeed, when we look at Figure 24 we can see that particularly at the urban fringes of Manchester high direct and indirect GHG emissions go hand-‐in-‐hand with low population densities. Only in very few we find low emissions in high density areas or a low footprint in low density areas. However, Figure 25 suggests that this could simply be due to the fact that in Manchester people with higher incomes have a preference to live in lower density areas. There are many other factors in the structure and socio-‐economic profile of the areas, which need to be explored to develop a better understanding what determines the global environmental impacts of urban life.
Figure 24 -‐ Densities and GHG emissions for the city of
Manchester
Figure 25 -‐ Income and CO2 emissions for the city of
Manchester
4.2.2 Quick data validation attempt Given our geodemographic approach to downscale CO2 and GHG emissions to the local scale based on expenditure clusters of lifestyle groups in a particular area, the suspicion remains that particularly the close relationship between income and emissions is a reflection of the methodology applied. Even though local information has been introduced as much as possible to improve the robustness of the estimates it is unclear whether these efforts to obtain a sufficient level of data quality. In this Section we use detailed domestic electricity and gas consumption data based on actual energy metering information in order to shed some light on the issue.
54 | P a g e
Figure 26 Relationship between average domestic electricity and gas consumption at middle layer super
output area in England and household income
Figure 27 -‐ Relationship between average domestic
electricity at middle layer super output area in England and household income
Looking at more detailed MLSOA level data the relationship between average weekly household income and domestic gas and electricity consumption is relatively weak. At a given income level domestic energy consumption can easily vary by a factor of two or more. The relationship gets slightly stronger when we focus on electricity consumption only. This could imply that the amount of gas used by household is more dependent on other factors such as the type of building (and therefore also often the area), the local climate etc..
In fact, when we look into individual cities as whole we find that the relationship between income and domestic energy consumption becomes much stronger as shown in Figure 28 and Figure 29 for Liverpool and Manchester. This could point towards the existence of structural determinants that might strongly influence domestic energy consumption levels for the city as a whole. Their identification is an interesting research question in itself.
Figure 28 -‐ Relationship between domestic energy
consumption and average weekly household income at MLSOA level for the city of Liverpool
Figure 29 -‐ Relationship between domestic energy
consumption and average weekly household income at MLSOA level for the city of Manchester
However, what does this imply for the downscaled emission estimates presented earlier? There seems to be a relationship between income and energy consumption and CO2 emissions. Clearly the carbon footprint evidence shows a much closer relationship to income across local authorities then this energy consumption data. However, the estimates also go far beyond housing related CO2
55 | P a g e
emissions. In fact, it can be shown that the relationship between direct and indirect CO2 emissions from housing and average weekly household income is far more spurious than in other consumption areas when we analyse at the local authority level. Therefore, the energy data here does at least not seem to invalidate the more comprehensive carbon footprint estimates and it appears worthwhile to put further efforts into finding viable options for downscaling emissions to small spatial scales.
4.2.3 Implications Given the difficulties in providing a comprehensive database for quantifying urban metabolism, it seems that a lot of different datasets will need to be used in order to shed light on the different research questions posed within the IUME programme. In this Section we have provided evidence for CO2 emissions of cities, which goes beyond what has been proposed for the indicator system (see Section 4.1) in at least three aspects:
A complete consumption based account has been provided, which covers all indirect CO2 emissions associated with consumption in cities;
Small area estimates of CO2 emissions have been provided not only for urban, but also rural areas;
CO2 emission estimates with a much higher spatial resolution have been provided.
Using very simple examples, we have demonstrated how this type of evidence can be used to answer specific policy questions and highlighted the higher levels of uncertainties, which are related to the construction of the dataset based on a downscaling methodology rooted in geo-‐demographics. While there can be severe limitations to such data, it should not be disregarded as there are little alternatives available right now. Instead we recommend:
The encouragement of result verification once other evidence becomes available. For example, there is similar municipal data for the UK (using a different dataset), Sweden and Norway. This data should be used to verify findings to a set of general questions (considering the specific country context);
The encouragement of research into downscaling methodologies. Many of the data gaps on the local level might be difficult to fill and are likely to remain open at list in the medium term. The development of downscaling methodologies, which use the best available information for their imputation efforts, could be a promising research avenue that has not been explored comprehensively. In fact, the importance of downscaling is increasingly becoming recognised in the literature on climate change adaptation (Hallegatte et al. 2008).
56 | P a g e
5 Discussion In this report we have developed a conceptual framework for the quantification of urban metabolism. We have then presented two empirical applications. First, we have established a comprehensive indicator system using the administrative boundaries of cities as spatial delineation. We have provided a quantification of this indicator system for the cities of Barcelona, Freiburg, Lille, Malmo and Sofia, and have outlined key avenues of analysis. Second, we have used UK specific data in order to address some of the shortcomings of the proposed indicator system: most importantly the importance of having comprehensive consumption based data with a higher degree of spatial granularity available to understand issues associated with urbanisation and sprawl.
In this Section we critically discuss the proposed urban metabolism concept and its operationalisation along the following four dimensions:
Urban metabolism as a systems approach; Linking urban metabolism to eco-‐system service provision; Importance of urban delineation approaches and spatial resolution for understanding urban
systems; Data availabilities & Data quality;
5.1.1 Taking a systems approach The main appeal of the urban metabolism concept is that it provides a systems approach to the analysis of urban areas. This entails depicturing all metabolic inflows and outflows associated with a
considerable environmental pressures at multiple scales and across environmental media and can indirectly influence ecosystems in very distant regions of the world as captured in the notion of a
The first challenge of this project is whether and how the systemic features of the metabolism concept can be captured at the city level based on publicly available data. In general, taking a systems approach is not a methodological but a data challenge. If we had perfect information, there would be methodologies available to correctly calculate the system wide environmental pressures imposed by a given vector of final consumption. However, having full information on physical flows triggered directly and indirectly throughout the global supply chain of goods and services consumed in a(n) (urban) area is impossible. In fact, we know from the life cycle assessment literature that this is even not the case for a single product (see Minx et al. 2008). Hence, methodological discussions are not a constitutional challenge associated with the calculation of system-‐wide environmental pressures, but the result of the imperfect data situation.
However, such general problems associated with system approaches might not be of major importance in the context of discussions on urban metabolism. Urban metabolism is interested in the system-‐wide environmental pressures associated with the consumption of larger bundles of goods and services (i.e. we are interested in all consumption from food to electricity to beauty products to transport demands). The literature shows that uncertainty reduces as the level of aggregation increases as errors tend to cancel when the consumption of a number of different products is assessed (see Bullard and Sebald 1977; Wiedmann et al. 2008; Lenzen et al. 2010). Therefore, given the availability of information on final consumption within a particular area, we can
57 | P a g e
expect to obtain reasonable top-‐level estimates for cities for example, based on environmental input-‐output analysis.
However, the problem associated with metabolism studies is that data on consumption of products, materials or energy is often not easily available (see Kennedy et al. 2007). This general data problem at the city scale is aggravated by the requirement of this particular project that data should be obtained for a large number of cities across Europe and be taken from regularly updated, publicly available data sources. This implies the need for a selective description of the urban metabolism focussing on key physical flows. In fact, certain consumption activities might not even be relevant for understanding urban systems, because they are not determined by the particular way how life is organised in a city or the available infrastructure. From this perspective the selection process might not even necessarily need to jeopardise a comprehensive understanding of urban systems, if we can find information in key consumption areas.
In terms of the proposed indicator system we have put an emphasis on energy, greenhouse gas emissions, water, land use and waste. Moreover, we have included some local air pollutants. From a functional perspective the areas of housing and transport are highlighted. The construction of system-‐wide indicators turned out to be difficult throughout. In the end we only included two indicators on direct water and energy consumption respectively. These indicators do not reflect the full energy and water requirements systems approach. We further propose a more comprehensive indicator on direct and indirect CO2 even though there is currently not sufficient public data available to calculate this indicator for a larger set of cities on a regular basis. Clearly, even though this is far from being perfect, it is a first step into the direction of a systems approach. In fact, we show that based on this information relevant analysis can be undertaken, which is largely missing today. Overall, the implementation of the proposed indicator system still seems worthwhile pursuing.
The second part of the analysis provided more comprehensive consumption based estimates of CO2 emissions for local areas in the UK. While such data has great appeal for urban metabolism research, it is difficult to perceive that such data will be available for a wider set of European countries in the near future. Moreover, much of the underlying data is estimated based on downscaling methodologies and uncertainties can be generally expected to be higher.
Resource Institute (WBCSD and WRI 2004). This terminology has been applied in the city context, for example, in the International Local Government GHG Emissions Analysis Protocol (ICLEI 2009).
Scope 1 emissions comprise all emissions released from the city territory; Scope 2 emissions are indirect emissions from the consumption of electricity on the city territory; Scope 3 emissions are all other indirect emissions activities.
Scope 2 and 3 emissions include emission sources outside the city territory.
Box 5 -‐ Emission scopes as defined in the GHG Protocol
58 | P a g e
On a more general level what is the best methodological approach to deal with the various data limitations in the future and calculate the system wide environmental pressures associated with a
can be found: The first is a top-‐down approach based on environmentally extended input-‐output analysis. This approach attempts to explicitly model the entire global production system and link it to a vector of local consumption activities (Druckman et al. 2008; Minx et al. 2009). One advantage is that it is easier to establish estimates of system wide environmental pressures from all consumption activities as they can work with less specific (sector or meso-‐level) information. Other methodologies such as material flow (Brunner and Rechberger 2004) or life cycle analysis (BSI 2006) start bottom-‐up from individual materials or products and try to calculate system-‐wide impacts by tracing individual supply chains. These approaches have the advantage that they provide more precise estimates for particular materials, products or technologies.11 Recently also the role of hybrid methodologies combining information from input-‐output and life cycle analysis have been proposed in the city context (Ramaswami et al. 2008; Larsen and Hertwich 2009; Hillman and Ramaswami 2010).
So which route is most promising to follow? For the development of the proposed indicator system, material flow and life cycle assessment based approaches are certainly better suited. They can be used to increase the scope of the consumption based indicators in a stepwise process (see Box 5). For example, in the case of the consumption based CO2 indicator it is straightforward to move from scope 1 to scope 2 emissions if data on energy consumption is available. Later other scope 3 elements should be added once new data sources emerge.
Input-‐output based approaches are worthwhile pursuing, if it is possible to establish local consumption expenditure vectors at the city level. One viable option therefore would be to scope possibilities for using national (and regional and local if available) consumer expenditure surveys across Europe for deriving this data. Clearly, this would at best allow the establishment of comprehensive consumption based resource and emission accounts for larger cities. The second option is to look at available downscaling methods and their implementation across Europe, but this is likely to require substantial resources. We will further comment on this point later on.
5.1.2 Linking to eco-‐system services and aspects of environmental quality The second major challenge of this project is to establish whether and how the urban metabolism concept can be related to ecosystem services using a simple indicator approach. An initial concept has been proposed and first steps have been taken in terms of a practical implementation. Even though it might be too early to judge the details of the proposed approach right now due to the on-‐going indicator construction efforts, at least some general aspects can be discussed already now.
First, even though we have highlighted in our concept the need to establish direct links to resource availabilities and sink capacities we have not or only very partially operationalised these aspects in our indicator system. For example, we only report per capita CO2 emissions and do not relate this to estimated sink capacities associated with some (arbitrary) agreed political target such as limiting
11 This is a simplification. However, it is not the ambition here to provide a detailed methodological discussion. Other studies have done this previously: e.g. Femia, A. and S. Moll (2005). Use of MFA-‐Related Family of Tools In Environmental Policy-‐Making -‐ Overview of Possibilities, Limitations and Existing Examples of Application In Practice. Copenhagen, European Environment Agency, European Topic Centre on Waste and Material Flows.
59 | P a g e
global warming to two degrees, which has been adopted by more than 100 countries so far (Meinshausen et al. 2009). However, we have discussed using the example of water how this is perceivable in the medium term. Similarly, we have not established a qualitative link between metabolic flows and eco-‐system functioning right now.
Second, in our proposed indicator system there is no causal relationship between changes in a particular ecosystem quality and the development of some metabolic flows. However, information about the state of an ecosystem say a freshwater body might provide some indication whether high levels of water consumption or the release of untreated waste water might potentially jeopardise the environmental integrity of this freshwater ecosystem. In this sense there are similarities to the pressure-‐driving force-‐state-‐response approach by the Organisation for Economic Co-‐operation and Development (OECD).
The task of establishing links between urban metabolic flows and eco-‐system services becomes even more complex once we also take into account all the indirect flow components. Recalling that we are dealing with global system boundaries, these indirect flows might occur anywhere in the world. This makes an assessment of any environmental pressures caused increasingly difficult particularly for regional and local pollutants. For example, coffee consumed in cities requires a lot water in its production. However, the effects of coffee production on local ecosystems depends on many regional and local determinants including the water scarcity of the region, the water pollution from coffee production etc..
Regardless of the complexity of the task, we might still be able to extract some relevant information once we introduce some simplifications. Figure 30 shows greenhouse gas emissions associated with meat imports to the UK (size of the cows) together with information on change in the forest cover in the UK. It is easy in such a graph to identify which part of the metabolism might put ecosystem services provision at risk in the long-‐run. The UK, for example, imports a considerable share of its beef from Brazil, where deforestation activities are causing not only a large amount of land-‐use change related CO2 emissions, but also biodiversity loss and change hydrological conditions among others (see Steinfeld et al. 2006; Steinfeld and Wassenaar 2007). Hence, describing ecosystem threats with such complexities associated with the system nature of the metabolism concept.
However, given these complexities it is key to consider the relevance of particular environmental issues in the urban context. The environmental consequences of cattle farming in Amazonia related to beef consumption in a particular city might be negligible even though the aggregate impacts of beef consumption in cities around the globe might be huge. Moreover, particular consumption activities might only be worthwhile considering in an (prioritizing) urban metabolism context, if a) we can show that it is a particular feature of an urban lifestyle; b) there are unique opportunities for addressing it at the local level, or, c) there are considerable overlaps with other important urban policies. It is therefore key for future research to find out which part of consumption patterns have local, regional and national characteristics. Only relevant local aspects of consumption should be dealt with at the local level.
60 | P a g e
=
Figure 30 -‐ Greenhouse has emissions from meat consumption and deforestation
Linking ecosystem services to the metabolism of cities on the conceptual level is an important issue. In the empirical application we have undertaken an initial step into this direction. An institution such as the European Environment Agency seems adequately placed with its large expertise across environmental topics to develop the approach further. One of the biggest challenges in this context might be whether it is possible to link existing indicators of ecosystem quality and functioning, to individual cities. However, it is clear that a simple indicator system can at best be an early warning system for potential stress imposed on ecosystem by the metabolism of cities. In order to understand causal relationships between metabolic flows and ecosystem services, only models can provide the required insights (e.g. Alberti 1999; Alberti et al. 2003; Alberti 2005).
5.1.3 Urban drivers and patterns The third major challenge in the context of this project is to establish whether and how changes/differences in the metabolism of cities might be related to changes in a set of drivers such as the lifestyle of people living in the city, temperature variations/ differences, price levels etc. or urban patterns including the spatial configuration of cities, their particular shape and development patterns. Linking urban metabolism conceptually to urban drivers and patterns is a fundamental requirement for addressing this issue and understanding metabolic flows in the context of urbanisation processes, sprawl etc..
We have proposed a variety of indicators to capture some of key determinants of urban patterns such as city size, urban form, transportation network or building stock. An important limitation associated with the proposed indicator system is that it currently works with administrative boundaries at the city scale. From a morphological perspective such an urban delineation is arbitrary. This can render comparisons of cities meaningless. For example, consider that within the administrative boundaries of a city is a highly densified urban area and the large remaining part is completely made out of forest. Any simple population density indicator at this spatial scale would be
61 | P a g e
highly distorted and fail to represent the structure of the urban area contained. For the example of Madrid, in Figure 31, we can further see that the urbanised might stretch beyond the territorial boundaries. Relevant parts of the urban area might be left out based on an administrative delineation.
Figure 31 -‐ Relationship between a morphological and adminstrative delineation of Madrid (Fons-‐Esteve, 2008)
We have opted for a compromise in our indicator system. Given the general unavailability of information on metabolic flows for functional and morphological delineations of urban areas, we focus on the administrative area only. This means that we neglect all other administrative areas containing the remaining parts of the morphological delineation of a city. Instead we use more detailed information in order to describe the specific pattern of land use and form on the administrative territory. Coming back to our example, we would not calculate the population density for the administrative area as a whole, but only for the sealed urban area (or area above a particular threshold of soil sealing) on the administrative territory of the city. Overall, this seems to be a pragmatic and potentially fruitful approach to shed light on the relationship between urban form and metabolic flow whilst remaining with the urban delineation where most physical data is available. The proposed metrics should be reviewed by experts and calculated on a regular basis e.g. in the context of the integrated urban monitoring project. Ways to downscale information on metabolic flows in a way that the data can be related to a morphological delineation of urban areas is discussed below.
5.1.4 Urban system and spatial resolution The proposed indicator system uses an administrative definition of the urban system as most urban environmental data is collected at this particular scale. Also recent data collection initiatives such as the Covenant of Mayors work with administrative boundaries (Covenant of Mayors 2010). In the previous Section we have already discussed the problems associated with such an urban delineation when we try to understand the relationship between urban patterns and metabolic flows associated
62 | P a g e
with the urban systems. However, we have also highlighted that the proposed approach should be in
Europe, which can be studied already at this spatial level. For example, it is intriguing to look beyond a specific city and analyse the general properties of European cities with a particular focus on resource consumption. How the size of cities relates to infrastructure investment and electricity or water consumption might provide important insights into the drivers behind these relationships and opportunities for targeted policy interventions (Bettencourt et al. 2007). In order to summarise some metabolic implications from urbanisation processes, we propose to introduce an aggregate indicator table for urban systems that compiles diverse scaling relationships between population size and key environmental, infrastructural and social factors as shown in Table 11.
However, the high level indicator system might be limited when we try to understand issues such as urbanisation or sprawl, which are diffuse and work at much smaller scales. There is, for example, the risk that we cannot easily identify the consequences of sprawl at such a high level of spatial aggregation or disentangle the effects of sprawl from other forces, which are at work simultaneously. The use of more detailed information should be encouraged where available (e.g. UK, Sweden, Norway) -‐ to study the metabolic implications of urbanisation processes in more detail.
Given the wide unavailability of socio-‐economic and environmental information there is a great need for exploring ways to downscale information. In the area of geodemographics, for example, a whole industry has developed which makes extensive use of such techniques for commercial purposes (Harris et al. 2005). It should therefore be explored whether such techniques deliver useful information for the urban research agenda in Europe, whether they could be applied across Europe and what this would entail. In the second half of the empirical part we have undertaken a first, simple exploration of such data for the analysis of drivers and patterns associated with metabolic flows triggered by consumption activities in five UK cities. Given that geo-‐demographic data systems are built from census information, this could be a timely initiative given that 2011 is the next census year in most European countries. However, geodemographics only provide one technique for downscaling information and other avenues should be explored as well.
5.1.5 Data availability & data sources The general data availability for populating the proposed indicator system for urban metabolism is reasonable. The vast majority of information was sourced from public databases, which are easily accessible via the internet. Admittedly, the indicator choice was partially determined by data availabilities given the task to choose a pragmatic approach, which works already today. However, regardless of the fact that we could not always choose the first best indicator option, this pragmatic approach does not seem to jeopardise the system as a whole. Moreover, indicators can be replaced instantaneously once new data is available.
Most variables are sourced from the urban audit the only comprehensive city level database in Europe with data across a wide range of themes covering demography, social and economic aspects, civic involvement, training and education, environment, travel and transport, information society as well as culture and recreation. Where possible we (recommend to) complement or substitute variables from the urban audit with information from other sources. This is particularly important as the consistency and robustness of urban audit variables can be limited. For example, it makes sense to take information on land-‐use and land cover from the Urban Atlas or CORINE, soil sealing data
63 | P a g e
from HR Soil Sealing or waste water data from the urban waste water treatment directive or WISE. The fact that urban audit data is not always available for a common year across cities is a minor drawback.
It is unfortunate that data availability is worst the central component with regard to the urban metabolism concept. The chosen approach has been very pragmatic in dealing with the systemic features of the metabolism concept (i.e. all metabolic flows
and climate change, land as well as waste only, on the one hand, and largely leaving out indirect metabolic flows required in the production of goods and services consumed in an urban area.
However, there still remain considerable data gaps. First, it is close to impossible to find any data from public data sources for a large range of cities, which allow the calculation of consumption based indicators in any of the thematic areas (energy & climate change, water, land, waste). Any viable process to derive consumption-‐based CO2 emission estimates for a larger set of cities will a) require some fundamental data developments ; and/or b) need to take a step-‐wise approach if results are to be expected in the not too far future and no major financial resources are to be spent. Given the importance of consumption based emission estimates for any framework for urban metabolism we recommend to develop this area actively.
Second, the current availability of any energy and GHG emission data is a general problem. Therefore, a key area of the indicator system cannot be easily populated. In fact, a larger implementation of this system does not seem worthwhile until this data gap is closed. The various on-‐going initiatives to standardise estimation methodologies and compile data are hoped to overcome this data shortage within the next six to twelve month. It is of utmost importance to encourage public accessibility of this data. Once available the data set will provide a decent starting point for analysis from purely publicly available data which is in scope comparable to what other projects have done based on expensive surveys.
5.1.6 Comparability and uncertainties It is difficult to provide a qualified judgement on data quality, uncertainties and comparability as there is often very little information available. In terms of the urban audit data two things are noteworthy. First, the organisational structure for the urban audit was set-‐up under the lead of Eurostat to foster a given quality of urban statistics (see, Eurostat 2004). Second, Eurostat provides ranges in which indicator estimates should fall. This enables users at least to undertake manual data checks (Eurostat 2004). For our test cities there we no deviations from these ranges. Overall, we should therefore expect the data to be of sufficient quality and comparability for our purposes here.
Land-‐use and land-‐cover statistics are even less problematic as long as they are derived from one consistent data source. It is therefore a recommendation to re-‐calculate all land-‐use and land-‐cover metrics from the dataset of choice (probably urban atlas) when the data system is established for a larger set of cities.
Little can be said about CO2 and energy statistics as it remains unclear where it will be ultimately sourced from. However, transparency and methodological consistency should benefit from the data development and standardisation efforts going on in this area right now. Before requesting cities to provide energy and CO2 emission data as part of a sustainable energy action plan (SEAP), the
64 | P a g e
Covenant of Mayors, for example, established a common methodology and reporting guidelines. This in combination with subsequent testing procedures should ensure a basic level of consistency and transparency, which should make the data suitable for the purposes here.
One important question is to what extent a production (or source) based accounting framework provides comparable data at the city level at all. For example, a large coal power plant on the city territory can mean that CO2 appear high even though only a fraction of this electricity is consumed by city residents. In other words, a large share of the emissions belong to the metabolism of other areas. From this perspective comparability increases when we move towards a consumption based account. It seems like a fundamental requirement that at least CO2 emissions from electricity generation should be assigned to cities according to use even though this idea could be extended to all economic activities associated with goods and services imported to or exported from cities. The carbon footprint data provided in this report for the five test cities is consistent in the way how it has been estimated. However, the high levels of uncertainty associated with the raw data causes concerns and require large verification and data collection (for improvements) efforts.
65 | P a g e
7 Summary and recommendations In this report we have developed a concept for the quantification of urban metabolism and proposed a pragmatic implementation based on publicly available data. Throughout the report we have highlighted the importance of going beyond a purely metabolic assessment and proposed three extensions of the standard urban metabolism concept:
Linking the urban metabolism to environmental pressures and aspects of environmental quality at multiple scales;
Linking urban metabolism to urban drivers, patterns and lifestyles; Linking urban metabolism to aspects of quality of life.
Accordingly, we propose a pragmatic indicator framework that has indicators of urban (metabolic) flows at its centre, but complements these with indicators characterising the physical structure of the city (urban patterns), the socio-‐economic drivers at work including lifestyle descriptors (urban drivers) as well as the quality of life in the city (urban quality). Thematically the indicator framework puts an emphasis on aspects of resource productivity and focuses on a limited set of metabolic flows: energy carriers, water, solid waste, emissions to water, emissions to air and land-‐use. In order to address the wide scope of questions identified under this service contract, we chose a relatively large number of indicators derived from a variety of data sources (including urban audit, urban atlas, Corine etc.). Due to the difficulties in dealing with such a wealth of information we designed a headline indicator set consisting of 14 indicators (see Figure 12), which summarise information directly relevant to key areas of the Aalborg commitment and the strategy of the sustainable use of resources.
In terms of the indicator design we have mainly opted for indicators that describe levels of resource/pollution flows (expressed as efficiencies) rather than changes over time. This is motivated by the comparative scope of this projects and its emphasis on the understanding of the relationship between metabolic flows and its underlying socio-‐economic and infrastructural drivers.12 However, an understanding of these relationships is generally difficult solely based on such average indicator data. We therefore propose two analytical extensions, which focus on relational measures such as elasticities: first, in order to understand better metabolic implications associated with urban growth, we propose the estimation of a set of scaling relationships focussing on aspects of resource use (see, Bettencourt et al. 2007). Second, in order to identify the marginal relationship between metabolic flows and urban drivers, urban patterns and urban quality we suggest the performance of regression studies based on the cross-‐sectional provided by the indicator set.
Figure 32 identifies the potential position of the urban metabolism framework as the systems approach within the IUME activities. This highlights a crucial point: the three pillars of IUME (data, system and questions) are mutually dependent. Success in obtaining answers will therefore depend on close collaboration across the pillars.
12 Further note that performance indicators, which measure changes over time, would be difficult to construct anyway in the light of data availabilities.
66 | P a g e
Figure 32 -‐ The urban metabolism framework in context of the Integrated Urban Monitoring in Europe (IUME) activities
Overall, we believe that the indicator set provides information comparable to what has been put forward by other city level. However, instead of using expensive survey techniques, this indicator set uses only publicly available data. While there might be disadvantages in the definition and consistency of indicators, we believe that this approach is worthwhile pursuing for at least two reasons:
It enables regular updating at almost no costs; We expect the data to be of sufficient quality to identify general trends in urbanisation
across Europe the main intention behind this service contract.
However, the discussion section above highlights the various challenges posed by the pragmatic approach taken in the implementation of the urban metabolism concept. Some of the aspects can therefore not be dealt with right now or only in a simplified manner. Below we provide some recommendation about actions in the short, medium and long term to reap the benefits of a systems to the understanding of cities in Europe:
1. Implement and analyse the pragmatic indicator framework for urban metabolism:
Initially, the proposed concept and indicator system should be reviewed, adjusted and then be implemented for a larger set of cities. Given the current unavailability of CO2 and energy data for a larger sample of cities, we recommend to wait with such an effort until the first set of SEAP data is published (see Data gaps and development section below) by the Covenant of Mayors (2010). We expect the first set of estimates to comprise at least 120 to 150 cities. The intersection with the urban audit sample should provide a sufficiently large testbed. This work could be carried out as part of the IUME initiative, within the EEA or under a small service contract. The work should start from the key analytical options identified (headline indicators, scaling relationships and multivariate regression) and evaluate the insights that can be gained for a pre-‐defined set of key questions.
67 | P a g e
2. Establish a working group for the development of the urban metabolism concept and indicator framework:
Enabling learning processes and institutionalising development processes are key, if a serious attempt is made to establish the metabolism approach as a core component of IUME and the urban ecosystem work at the EEA. The focus of this project on pragmatism emphasises the need for further development as there is still a large gap between the conceptual ideas put forward in this report and their implementation: closer links of the urban metabolism to aspects of local, regional and global environmental quality or environmental pressure are one example. At the same time there is a considerable momentum in urban research across Europe and we believe that this could lead to a swift development of urban metabolism concept and framework.
We therefore propose the establishment of a working group for the development of the urban metabolism concept and indicator framework. This working group could be embedded into the on-‐going IUME activities, be headed by a representative of the EEA and involve existing IUME members as well as experts from the field of urban metabolism. Experts could involve key representatives
th framework programme, members of other EC institution (DG Regio, Covenant of Mayors etc.) and other representatives from the field. This working group could meet annually or bi-‐annually and review the urban metabolism concept and its implementation. This would include the identification of new data availabilities (e.g. WISE, Urban Atlas, Corine, etc.) and resulting opportunities for improving the indicator framework as well as reports on relevant research activities across Europe and recommendations on future developments.
3. Derive additional insights from case study evidence:
The proposed pragmatic approach to quantifying urban metabolism will not be able to provide insights into all questions of interest outlined by the EEA. For example, carbon, energy or water footprint data is required for fully assessing questions directed at the relationship between lifestyles, urban form and environmental impact. Similarly, some of the questions related to sprawl and densification key issues in the discussion would benefit from environmental data with a higher spatial resolution. Such data is not available for a larger set of European Cities and it is not perceivable that comprehensive and consistent accounts can be established without a substantial investment of financial resources. However, there is some evidence on energy and carbon footprints of municipalities, for example, in the UK13, Sweden and Norway14 and some domestic energy use/ CO2 accounts for the UK and some federal state in Germany with a higher spatial resolution15. Pragmatism suggests to use such data for providing some answers to questions, which cannot not be addressed with the proposed indicator framework at the moment.
4. Encourage key data developments
The urban audit has improved the general data availability for European cities substantially. The work of the EEA and its topic centres have complemented the urban audit data to some extent with environmental data, but considerable gaps remain. In general, the unavailability of environmental
13 See, http://www.resource-‐accounting.org.uk/downloads/?page=downloads) 14 See, http://www.klimakost.no/uk/ 15 See, http://www.decc.gov.uk/en/content/cms/statistics/regional/regional.aspx, http://www.statistik.baden-‐wuerttemberg.de/SRDB/home.asp?H=UmweltVerkehr&U=02
68 | P a g e
data is unequally distributed across the four dimensions of the proposed urban metabolism concept. The most considerable gaps are in the area of metabolic flow data (see Kennedy et al. 2007) the central element of the urban metabolism concept. However, this report also suggests that instead of engaging into additional costly data collection efforts, some smaller targeted action might be sufficient initially for a basic implementation of the framework. In the middle and long-‐term in-‐house data developments at the EEA and adjustments of on-‐going data collection initiatives such as urban audit could provide a sufficient basis for the development of the proposed urban metabolism concept and its proposed implementation. With regard to data development, we see the following key actions:
a) Encourage the publication of energy and CO2 (and GHG) data The most substantial data gap is the absence of energy and GHG emission accounts for a larger sample of European cities. Even though the urban audit asks requests energy and CO2/GHG related indicators, the response rate was so low that the data was never published. Filling this data gap is fundamental for the initial implementation of the proposed pragmatic indicator framework for a wider range of European cities. However, over the last few years a variety of initiatives have mushroomed, which help cities to produce energy and GHG emission statistics based on some defined set of guidelines. Government GHG Emissio (ICLEI 2009) as well as the Covenant of Mayors and
(Covenant of Mayors 2010). Hence, rather than collecting more data, the EEA should encourage the publication and easy access of data from these initiatives.16 Even if data is not made publicly available, these initiatives should ensure a decent stock of data for the next urban audit round. There could be a need for the reconciliation and integration of data published by different initiatives to ensure consistency and comparability. Consistent scope 2 emission accounts (see Box 5) should be established for CO2 from the energy data to enable more meaningful comparisons of cities (Kennedy et al. 2009; Kennedy et al. 2010).
b) Continue on-‐going in-‐house data developments in the area of water and land-‐use and ensure EEA and its topic centres are already producing relevant data for the proposed urban metabolism indicator framework particularly in the area of land-‐use (CORINE, Urban Atlas, Soil sealing etc.), but also in the area of water (e.g. WISE). On-‐going development processes could be further targeted to improve the quality of the indicator framework. In terms of land-‐use and land there are developments of immediate interest. First, a set of suitable landscape metrics (see Huang et al. 2007; Schwarz 2010) should be calculated from the Urban Atlas (i.e. for all Urban Atlas cities) to characterise physical structure and form of the cities. In this context, indicators UP7-‐UP9 should be reviewed. Second, indicator UF9 (increase in soil sealing by type of converted land) should be derived from Urban Atlas data and complemented and/or replaced by other indicators of urban sprawl and its potential pressures on the local environment.
The current water related indicators largely fail to establish clear links between urban water use and aspects of local/regional environmental pressures as well as environmental quality. While we have proposed an water related indicator set in Box 2, we propose a step-‐wise development as new data becomes available in the context of WISE. Emphasis should be given to defining source side indicators, which define aspects of water scarcity in the water body of abstraction (e.g. 16 For example, in the case of the SEAP data no decision has been made yet by the European Commission whether and at which level of detail data will be published.
69 | P a g e
groundwater, river flow, size of water body) or the (change in the) area from where the city sources its water. On the sink side, information is required about the (change in) water quality of the water bodies in which the water is released after use.
c) Transport and housing related data From a functional perspective the areas of housing (or buildings in general) and transport are of central importance for understanding urban metabolism (Weisz and Steinberger 2010). More detailed and more comprehensive data would be desirable in these two areas. With respect to transport there is currently a study under way on urban mobility in Europe by DG TREN, where data gaps will be identified and a roadmap will be designed for filling these. This report reemphasises the importance of this process and the need for a tangible output in terms of an improvement of the database on urban mobility.
In the area of housing more comprehensive statistics in terms of spatial coverage (beyond cities) and the spatial aggregation (more detailed information). Domestic electricity and gas use is generally metered across Europe. The UK the government requests power companies to supply electricity and gas meter data of their customers and publishes this information at a low level of spatial aggregation for the entire country (see footnote 15). This provides important information particularly with regard to differences in the energy requirements from urban, sub-‐urban and rural life. It should be a long term goal to develop similar statistics for Europe as a whole.
d) Identify new key environmental variables for the next urban audit Even though the urban audit provides a rich reservoir of urban statistics, environmental information is limited. As mentioned before the most immediate data gaps are associated with energy and CO2, but the absence of data is rooted in a lack of response for these variables by the cities. Given on-‐going data collection efforts by cities over the last few years, we can expect the next urban audit to be more successful in this area. Energy related variables defined in Section 6.6 of the Urban Audit appear generally appropriate (even though information on the share of imported electricity and the type of generation on the city territory including renewable energy technologies would be helpful), while more detailed information on CO2 emissions (e.g. domestic/ commercial split, process/ energy related splits etc.) and other GHG emissions in Section 6.2 would be desirable acknowledging the possibility to construct important components of the CO2 from the energy variables (see Eurostat 2007; Eurostat 2009).
(urban audit variables EN3003V-‐EN3010V) urban audit variables could be improved by providing a split into domestic and commercial water consumption, additional information about the amount of water extraction on the city territory, leakage as well as the type of water treatment technology applied. Transport statistics suffer mainly from the absence of information on car traffic volumes and more detailed statistics of the public transport system and its use by residents would help. Hence, on a more general level, despite the absence of certain information it is often the split into a domestic and import component, domestic and commercial users and more specific technology descriptions, which are desirable.
A main difficulty with most of the urban audit variables as previously discussed -‐ is their exclusive availability at the city level only. The value of the data would be greatly increased, if environmental variables would also be provided at LUZ and Sub-‐City level. A careful evaluation how this can be achieved could greatly benefit urban metabolism research in Europe.
70 | P a g e
5. Future research needs
There is a large need for strengthening sustainability research at the city scale, which became obvious in the course of this project:
a) Downscaling methodologies: One area for future research are methodologies for downscaling data to levels of higher spatial resolution. A higher level of spatial resolution would be of great value to study links between urbanisation dynamics, sprawl and their environmental impacts in more detail. Moreover, downscaling methodologies could potentially make metabolic flow data available not only available for administrative geographies, but also for functional and morphological urban delineations. This would ease the transition between different urban delineations and would enable to apply appropriate delineations to research questions (Fons-‐Esteve et al. 2008). In Section 4.2 we have introduced geodemographics as one potential way of doing so. In the United Kingdom, for example, there is an open and free geodemographic system available, where data can be readily obtained from by researchers (for a relevant application in the area of urban metabolism, see Druckman et al. 2008). A similar system could be envisioned for Europe as a whole (Harris et al. 2005). However, there are many other methodologies for downscaling data and it is of importance to understand how they can be applied and which data sources (such as Airbase or EPER) they should combine for the downscaling process. There is already expertise within the EC17 and a general interest in downscaling likely to be found across different directorates of the EC including DG Environment and DG Regio. A scoping of requirements for data with a higher spatial resolution across European institutions could be a sensible first step followed by a review of methods for downscaling information.
b) City typologies with a focus on metabolic aspects: Another important step for understanding the relationship between aspects of urban quality, urban form, urban drivers and the physical metabolism of a city is the development of an urban typology. While a variety of studies have identified groups of cities based on a set of landscape metrics and socio-‐economic indicators (e.g. Schwarz 2010), the consideration of metabolic features has been neglected so far
Hinterland relationships and development as well as a particular set of drivers would define a characteristic metabolic flow pattern. The existence of such archetypes would not only improve the understanding of the metabolism of European cities, but also be of direct value for supporting urban policy development at the European level as well as environmental research in Europe.
c) Stocks: The proposed indicator framework for urban metabolism puts an emphasis on metabolic flows. Stocks are only covered in the most generic way, for example, as number and types of buildings or as the physical shape of the urban infrastructure. Even though this might provide some insights into how the form of a city might determine metabolic flows to some extent,18 it leaves out a series of other questions related to stocks such as stock characteristics (different technology mixes and
17 See http://edgar.jrc.ec.europa.eu/index.php 18 We could, for example, consider metabolic flows to be a function of stocks (e.g. technologies) and use behaviour in this line of reasoning.
71 | P a g e
performances) and their dynamics from changes over time. While there is room for improving the representation of stocks in the current framework over time, there is a need for a wider and more systematic consideration of the role of stocks in the metabolism of cities, which we perceive lies mainly outside the scope of this urban metabolism framework. The development of adequate data for larger sets of European cities and targeted research on the role of stocks such as energy, water, transport, housing or waste infrastructures is recommended.
d) Consumption based resource accounting It might take some time until comprehensive consumption based resource flow accounts are available for a larger set of European cities and can be used to extend the current indicator set. It is therefore even more important that related issues are addressed by urban research in the meantime. There are a great variety of issues that need to be tackled. So far, there are a few studies available analysing the carbon, energy or ecological footprint of cities (Rees and Wackernagel 1996; SEI et al. 2006; Stockholm Environment Institute 2007; Druckman et al. 2008; Ramaswami et al. 2008; Larsen and Hertwich 2009; Hillman and Ramaswami 2010; Minx et al. 2010; Parshall et al. 2010). Still, there is a limited understanding of what drives these global impacts of cities and at what scale they should be addressed. There is a need to undertake such an analysis for wider categories of resource flows including water, waste and materials. Furthermore, it would be intriguing to ask how consumption in European cities contributes to air pollution problems, say, in the Pearl River Delta. Essentially, this would mean to analyse interdependencies between cities (particularly in developed and developing countries) through production and consumption networks. More fundamentally, research is required to answer the question in which areas of consumption footprint evidence might be required at the city scale. For example, if food consumption patterns in urban and rural areas are not substantially different and the set of potential policy solutions are also very similar, this issue might be best dealt with at the national level. Such research will complement insights obtained from the proposed urban metabolism indicator framework and should be encouraged.
e) Dynamic models The proposed urban metabolism concept focuses on understanding aspects of urban metabolism through a descriptive indicator system and a few simple analytical tools. To gain further insights researcher need to understand the mechanisms that describe the relationship between metabolic flows and aspects of environmental quality, urban patterns, drivers or quality of life in cities. Ultimately the aim of urban metabolism research is to understand how cities can move from one metabolic state (e.g. uni-‐directional economy) to another (e.g. circular economy).This requires the application of dynamic models in a scenario approach (Alberti 1999). This avenue of research is not very developed even though first European projects such as SUME are making steps into this direction.19 Such research will be crucial to move from an understanding of the past to the evaluation and shaping of the future.
19 See, http://www.sume.at/project_downloads
72 | P a g e
8 Literature AEA Technology (2008). Local and Regional CO2 Emission Estimates for 2006-‐2006 for the UK. Report to the Department for Environment, Food and Rural Affairs. Oxford. Alberti, M. (1996). "Measuring urban sustainability." Environmental Impact Assessment Review 16(4-‐6): 381-‐424. Alberti, M. (1999). "Modeling the urban ecosystem: a conceptual framework." Environment and Planning B: Planning and Design 26(4): 605-‐630. Alberti, M. (1999). "Urban Patterns and Environmental Performance: What do we know?" Journal of Planning Education and Research 19: 151-‐163. Alberti, M. (2005). "The Effects of Urban Patterns on Ecosystem Functions." International Regional Science Review 28(2): 168. Alberti, M., J. M. Marzluff, E. Shulenberger, G. Bradley, C. Ryan and C. Zumbrunnen (2003). "Integrating Humans into Ecology: Opportunities and Challenges for Studying Urban Ecosystems." BioScience 53(12): 1169-‐1179. Ambiente Italia Research Institute (2003). European Common Indicators. Milano. Aunan, K., J. Fang, T. Hu, H. M. Seip and H. Vennemo (2006). "Climate Change and Air
-‐Benefits in China." Environmental Science & Technology 40(16): 4822-‐4829. Aunan, K., J. Fang, H. Vennemo, K. Oye and H. M. Seip (2004). "Co-‐benefits of climate policy-‐-‐lessons learned from a study in Shanxi, China." Energy Policy 32(4): 567-‐581. Bagliani, M., A. Galli, V. Niccolucci and N. Marchettini (2008). "Ecological footprint analysis applied to a sub-‐national area: The case of the Province of Siena (Italy)." Journal of Environmental Management 86(2): 354-‐364. Baiocchi, G. and J. C. Minx (2010). "Understanding Changes in the UK's CO2 Emissions: A Global Perspective." Environmental Science & Technology 44(4): 1177-‐1184. Baiocchi, G., J. C. Minx and K. Hubacek (2009). "The Impact of Social Factors and Consumer Behavior on CO2 emissions in the UK: a Panel Regression Based on Input-‐Output and Geo-‐demographic Consumer Segmentation Data." Journal of Industrial Ecology under revision. Baiocchi, G., J. C. Minx and K. Hubacek (2010). "The Impact of Social Factors and Consumer Behavior on CO2 emissions in the UK: a Panel Regression Based on Input-‐Output and Geo-‐demographic Consumer Segmentation Data." Journal of Industrial Ecology 14(1): 50-‐72. Barles, S. (2007). "Feeding the city: Food consumption and flow of nitrogen, Paris, 1801-‐1914." Science of The Total Environment 375(1-‐3): 48-‐58. Barrett, J. (1998). Sustainability Indicators and Ecological Footprints: the case of Guernsey, School of the Built Environment, Liverpool John Moores University, Liverpool.
73 | P a g e
Barrett, J., R. Birch, N. Cherrett and T. Wiedmann (2005). Reducing Wales' Ecological Footprint -‐ Main Report. York, WWF Cymru, Cardiff, UK. Barrett, J. and A. Scott. (2001). "An Ecological footprint of Liverpool: A Detailed Examination of Ecological Sustainability." Berrini, M. and L. Bono (2007). Urban Ecosystem Europe -‐ An Integrated Assessment on the Sustainability of 32 European Cities, Ambiente Italia. Bettencourt, L. M. A., J. Lobo, D. Helbing, C. Kühnert and G. B. West (2007). "Growth, innovation, scaling, and the pace of life in cities." Proceedings of the National Academy of Sciences 104(17): 7301-‐7306. Bicknell, K. B., R. J. Ball, R. Cullen and H. R. Bigsby (1998). "New methodology for the ecological footprint with an application to the New Zealand economy." Ecological Economics 27(2): 149-‐160. Brown, M. A., F. Southworth and A. Sarzynski (2008). Shrinking the Carbon Footprint of Metropolitan America. Browne, D., B. O'Regan and R. Moles (2009). "Assessment of total urban metabolism and metabolic inefficiency in an Irish city-‐region." Waste Management 29(10): 2765-‐2771. Brunner, P. H. (2007). "Reshaping Urban Metabolism." Journal of Industrial Ecology 11(2): 11-‐13. Brunner, P. H. and H. Rechberger (2004). Practical Handbook of Material Flow Analysis. Boca Raton, Florida, CRC Print, Lewis Publishers. BSI (2006). ISO 14044: Environmental Management -‐ Life Cycle Assessment -‐ Requirements and Guidelines. International Organisation for Standardisation. London, British Standards Institute. BS EN ISO 14044:2006. Bullard, C. W. and A. V. Sebald (1977). "Effects of parametric uncertainty and technological change on input-‐output models." Review of Economics and Statistics LIX: 75-‐81. Carballo Penela, A. and C. Sebastián Villasante (2008). "Applying physical input-‐output tables of energy to estimate the energy ecological footprint (EEF) of Galicia (NW Spain)." Energy Policy 36(3): 1148-‐1163. Covenant of Mayors (2010). How to Develop of Sustainable Energy Action Plan (SEAP) -‐ Guidebook. Brussels, Publication Office of the European Union. Dhakal, S. (2009). "Urban energy use and carbon emissions from cities in China and policy implications." Energy Policy In Press, Corrected Proof. Dodman, D. (2009). "Blaming cities for climate change? An analysis of urban greenhouse gas emissions inventories." Environment and Urbanization 21(1): 185-‐201. Druckman, A., P. Sinclair and T. Jackson (2008). "A geographically and socio-‐economically disaggregated local household consumption model for the UK." Journal of Cleaner Production 16(7): 870-‐880.
74 | P a g e
European Commission (2005). Thematic Strategy on the Sustainable Use of Natural Resources. Brussels, Belgium, European Commission. SEC(2005) 1683. European Environment Agency (1996). Europe's Environment: The Dobris Assessment. London. European Environment Agency (2009). Ensuring Quality of Life in Europe's Cities and Towns -‐ Tackling the Environmental Challenges Driven by European and Global Change. European Environment Agency Report. B. Georgi and R. Uhel. Copenhagen. Eurostat (2001). Economy-‐wide material flow accounts and derived indicators (Edition 2000). A methodological guide. Luxembourg, Eurostat; Office for Official Publications of the European Communities, Statistical Office of the European Communities (Ed.). Eurostat (2004). Urban Audit -‐ Methodological Handbook. Luxembourg, Office for the Publications of the European Communities. Eurostat (2007). Urban Audit Reference Guide Data 2003/04 -‐ 2007 edition. Eurostat Methodologies and Working Papers, Office for Official Publications of the European Communities. Eurostat (2009). European Regional and Urban Statistics Reference Guide (2009 edition). Eurostat Methodologies and Working Papers, Office for Official Publications of the European Communities. Femia, A. and S. Moll (2005). Use of MFA-‐Related Family of Tools In Environmental Policy-‐Making -‐ Overview of Possibilities, Limitations and Existing Examples of Application In Practice. Copenhagen, European Environment Agency, European Topic Centre on Waste and Material Flows. Fons-‐Esteve, J., S. Kleeschulte, L. Guerrieri and M. Falconi (2008). "A Framework for Integrated Urban Monitoring in Europe." Working Paper Universitat Autonoma de Barcelona. Frey, B. S. and A. Stutzer (2002). "What Can Economists Learn from Happiness Research?" Journal of Economic Literature 40(2): 402-‐435. GAP, SEI and Eco-‐Logica (2006). UK Schools Carbon Footprint Scoping Study. London Report by Global Action Plan, Stockholm Environment Institute and Eco-‐Logica Ltd for the Sustainable Development Commission. Hallegatte, S., F. Henriet and J. Corfee-‐Morlot (2008). "The Economics of Climate Change Impacts and Policy Benefits at City Scale: A Conceptual Framework." OECD Environment Work Papers 4. Haq, G., J. C. Minx, J. Whitelegg and A. Owen (2007). Greening the Greys: Climate Change and the Over 50's. York, Stockholm Environment Institute. Harris, R., P. Sleight and R. Webber (2005). Geodemographics, GIS and Neighborhood Targeting. Chichester, John Wiley & Sons Ltd. Hendriks, C., R. Obernosterer, uuml, D. ller, S. Kytzia, P. Baccini and P. H. Brunner (2000). "Material Flow Analysis: a tool to support environmental policy decision making. Case-‐studies on the city of Vienna and the Swiss lowlands." Local Environment 5: 311-‐328. Hertle, H. and K. Schaechtele (2009). Who's ahead? Climate Cities Benchmark in Japan, U.S. and Germany. Heidelberg, IFEU Institute.
75 | P a g e
Hillman, T. and A. Ramaswami (2010). "Greenhouse Gas Emission Footprints and Energy Use Benchmarks for Eight U.S. Cities." Environmental Science & Technology 44(6): 1902-‐1910. Holden, E. and I. T. Norland (2005). "Three Challenges for the Compact City as a Sustainable Urban Form: Household Consumption of Energy and Transport in Eight Residential Areas in the Greater Oslo Region." Urban Studies 42(12): 2145-‐2166. Holling, C. S. (1977). Adaptive environmental assessment and management. New York, Wiley. Hu, D., S.-‐l. Huang, Q. Feng, F. Li, J.-‐j. Zhao, Y.-‐h. Zhao and B.-‐n. Wang (2008). "Relationships between rapid urban development and the appropriation of ecosystems in Jiangyin City, Eastern China." Landscape and Urban Planning 87(3): 180-‐191. Huang, J., X. X. Lu and J. M. Sellers (2007). "A global comparative analysis of urban form: Applying spatial metrics and remote sensing." Landscape and Urban Planning 82(4): 184-‐197. Huang, S.-‐L. and W.-‐L. Hsu (2003). "Materials flow analysis and emergy evaluation of Taipei's urban construction." Landscape and Urban Planning 63(2): 61-‐74. Hubacek, K., D. Guan, J. Barrett and T. Wiedmann (2009). "Environmental implications of urbanization and lifestyle change in China: Ecological and Water Footprints." Journal of Cleaner Production 17(14): 1241-‐1248. ICLEI (2009). International Local Government GHG Emission Analysis Protocol: Version 1.0 October 2009. ICLEI and Carbon Disclosure Project (2008). Carbon Disclosure Project -‐ Cities Pilot Project 2008, New York. Jenerette, G. D., W. Wu, S. Goldsmith, W. A. Marussich and W. John Roach (2006). "Contrasting water footprints of cities in China and the United States." Ecological Economics 57(3): 346-‐358. Kennedy, C., J. Cuddihy and J. Engel-‐Yan (2007). "The Changing Metabolism of Cities." Journal of Industrial Ecology 11(2): 43-‐59. Kennedy, C., J. Steinberger, B. Gasson, Y. Hansen, T. Hillman, M. HavraÌ nek, D. Pataki, A. Phdungsilp, A. Ramaswami and G. V. Mendez (2009). "Greenhouse Gas Emissions from Global Cities." Environmental Science & Technology 43(19): 7297-‐7302. Kennedy, C., J. Steinberger, B. Gasson, Y. Hansen, T. Hillman, M. Havránek, D. Pataki, A. Phdungsilp, A. Ramaswami and G. V. Mendez "Methodology for inventorying greenhouse gas emissions from global cities." Energy Policy In Press, Corrected Proof. Kennedy, C., J. Steinberger, B. Gasson, Y. Hansen, T. Hillman, M. Havránek, D. Pataki, A. Phdungsilp, A. Ramaswami and G. V. Mendez (2010). "Methodology for inventorying greenhouse gas emissions from global cities." Energy Policy 38(9): 4828-‐4837. Kennedy, C. A., A. Ramaswami, S. Carney and S. Dhakal (2009). Greenhouse Gas Emission Baselines for Global Cities and Metropolitean Regions. Urban Research Symposium. Marseille.
76 | P a g e
Kenworthy, J. R. and F. B. Laube (1996). "Automobile dependence in cities: an international comparison of urban transport and land use patterns with implications for sustainability." Environmental Impact Assessment Review 16(4-‐6): 279-‐308. Keuning, S. J. (1994). "The SAM and beyond: open, SESAME!" Economic Systems Research 6(1): 21-‐50. Keuning, S. J., J. van Dalen and M. de Haan (1999). "The Netherlands 'NAMEA'; presentation, usage and future extensions." Structural Change and Economic Dynamics 10(1): 15-‐37. Larsen, H. N. and E. G. Hertwich (2009). "The case for consumption-‐based accounting of greenhouse gas emissions to promote local climate action." Environmental Science & Policy 12(7): 791-‐798. Lenzen, M., C. Dey and B. Foran (2004). "Energy requirements of Sydney households." Ecological Economics 49(3): 375-‐399. Lenzen, M. and G. M. Peters (2009). "How City Dwellers Affect Their Resource Hinterland -‐ A Spatial Impact Study of Australian Households." Journal of Industrial Ecology forthcoming. Lenzen, M., R. Wood and T. Wiedmann (2010). "UNCERTAINTY ANALYSIS FOR MULTI-‐REGION INPUT-‐OUTPUT MODELS: A CASE STUDY OF THE UK'S CARBON FOOTPRINT." Economic Systems Research 22(1): 43 -‐ 63. McDonald, G. and M. Patterson (2003). Ecological Footprints of New Zealand and its Regions. Wellington, New Zealand, Ministry for Environment New Zealand. McDonald, G. W. and M. G. Patterson (2004). "Ecological Footprints and interdependencies of New Zealand regions." Ecological Economics 50(1-‐2): 49. McDonald, G. W. and M. G. Patterson (2004). "Ecological Footprints and interdependencies of New Zealand regions." Ecological Economics 50(1-‐2): 49-‐67. Meinshausen, M., N. Meinshausen, W. Hare, S. C. B. Raper, K. Frieler, R. Knutti, D. J. Frame and M. R. Allen (2009). "Greenhouse-‐gas emission targets for limiting global warming to 2[thinsp][deg]C." Nature 458(7242): 1158-‐1162. Minx, J. (2008). Integrated Data Frameworks in Monetary, Physical and Time Units for Quantitative Sustainable Consumption Research. Environment Department. York, University of York. PhD. Minx, J. C., J. Barrett, K. Feng, K. Hubacek and T. Wiedmann (2010). "The Carbon Footprint of Cities." prepared for submission. Minx, J. C., V. Medinger and T. Ziegler (2009). Developing a pragmatic approach to assess urban environmental impact on Europe based on the metabolism concept: An initial review, Stockholm Environment Institute, Technische Universität Berlin. Minx, J. C., T. Wiedmann, J. Barrett and S. Suh (2008). Methods Review to Support the PAS Process for the Calculation of the Greenhouse Gas Emissions Embodied in Good and Services: Report to the UK Department for Environment, Food and Rural Affairs. York, Stockholm Environment Institue and University of Minnesota.
77 | P a g e
Minx, J. C., T. Wiedmann, R. Wood, G. Peters, M. Lenzen, A. Owen, K. Scott, J. Barrett, K. Hubacek, G. Baiocchi, A. Paul, E. Dawkins, J. Briggs, D. Guan, S. Suh and F. Ackerman (2009). "Input-‐Output Analysis and Carbon Footprinting: An overview of UK applications." Economic Systems Research 21(3): 187-‐216. Moran, D., M. Wackernagel, J. Kitzes, B. W. Heumann, D. Phan and S. Goldfinger (2007). Trading Spaces: Calculating Embodied Ecological Footprints in International Trade Using A Product Land Use Matrix (PLUM). International Ecological Footprint Conference, Cardiff, Wales, UK, ESRC BRASS Research Centre, Cardiff University. Muñiz, I. and A. Galindo (2005). "Urban form and the ecological footprint of commuting. The case of Barcelona." Ecological Economics 55(4): 499-‐514. Newman, P. (2006). "The environmental impact of cities." Environment and Urbanization 18(2): 275-‐295. Newman, P., R. Birrell, D. Homes, C. Mathers, P. Newton, G. Oakley, A. O'Connor, B. Walker, A. Spessa and D. Tait (1996). State of the Environment. Chapter 3 -‐ Human Settlements in Australia. R. Taylor. Melbourne, Department of Sport and Territories. Newman, P., J. Kenworthy and P. Vintila (1995). "Can we overcome automobile dependence? : Physical planning in an age of urban cynicism." Cities 12(1): 53-‐65. Newman, P. W. G. (1999). "Sustainability and cities: extending the metabolism model." Landscape and Urban Planning 44(4): 219-‐226. Newman, P. W. G. and J. R. Kenworthy (1989). Cities and Automobile Dependence: A Sourcebook. Aldershot, UK, Gower Technical. Newman, P. W. G. and J. R. Kenworthy (1996). "The land use-‐-‐transport connection : An overview." Land Use Policy 13(1): 1-‐22. Niza, S., L. Rosado and P. Ferrão (2009). "Urban Metabolism." Journal of Industrial Ecology 13(3): 384-‐405. Owen, A., A. Paul and J. Barrett (2008). Ashford's Footprint -‐ Now and in the Future. SEI Working Paper. York, Stockholm Environment Institute. Parshall, L., K. Gurney, S. A. Hammer, D. Mendoza, Y. Zhou and S. Geethakumar (2010). "Modeling energy consumption and CO2 emissions at the urban scale: Methodological challenges and insights from the United States." Energy Policy 38(9): 4765-‐4782. Parshall, L., S. Hammer and K. Gurney (2009). Energy Consumption and CO2 Emissions in Urban Countires in the United States With A Case Study of the New York Metropolitean Area. Fifth Urban Research Symposium. Marseille. Ramaswami, A., T. Hillman, B. Janson, M. Reiner and G. Thomas (2008). "A Demand-‐Centered, Hybrid Life-‐Cycle Methodology for City-‐Scale Greenhouse Gas Inventories." Environmental Science & Technology 42(17): 6455-‐6461.
78 | P a g e
Rauch, J. N. (2009). "Global mapping of Al, Cu, Fe, and Zn in-‐use stocks and in-‐ground resources." Proceedings of the National Academy of Sciences 106(45): 18920-‐18925. Rees, W. and M. Wackernagel (1996). "Urban Ecological Footprints: Why Cities Cannot be Sustainable -‐ And Why They are a Key to Sustainability." Environmental Impact Assessment Review 16: 223-‐248. Rees, W. and M. Wackernagel (1996). "Urban ecological footprints: why cities cannot be sustainable -‐ and why they are a key to sustainability." Environmental Impact Assessment Review 16(4-‐6): 223-‐248. Schwarz, N. (2010). "Urban form revisited-‐-‐Selecting indicators for characterising European cities." Landscape and Urban Planning 96(1): 29-‐47. Schwela, D., G. Hay, C. Huizenga, W.-‐J. Han, H. Fabian and M. Ajero, Eds. (2007). Urban Air Pollution in Asian Cities -‐ Status, Challenges and Management. London, Earthscan. Scotti, M., C. Bondavalli and A. Bodini (2009). "Ecological Footprint as a tool for local sustainability: The municipality of Piacenza (Italy) as a case study." Environmental Impact Assessment Review 29(1): 39-‐50. SEI (2008). Wales' Ecological Footprint -‐ Scenarios to 2020. SEI Working Paper. York, Stockholm Environment Institute. SEI, WWF and CURE (2006). Counting Consumption -‐ CO2 emissions, material flows and Ecological Footprint of the UK by region and devolved country. WWF-‐UK. Godalming, Surrey, UK, Published by WWF-‐UK. Stahmer, C. (2000). The Magic Triangle Of Input-‐Output Tables, Macerata, Italy, 13th International Conference on Input-‐Output Techniques, 21 -‐ 25 August 2000, Macerata, Italy. Steinfeld, H., P. Gerber, T. Wassenhaar, V. Castel, M. Rosales and C. de Haan (2006). Livestock's Long Shadow. Rome, Food and Agriculture Oranization of the United Nations. Steinfeld, H. and T. Wassenaar (2007). "The Role of Livestock Production in Carbon and Nitrogen Cycles." Annual Review of Environment and Resources 32(1): 271-‐294. Stockholm Environment Institute (2007). The Right Climate to Change: The Carbon Footprint of UK Local Authorities. Godalming, WWF-‐UK. Suter, A. (1991). Noise and Its Effetcs. Paper prepared for the Administrative Conference of the United States of America. Timmer, S. and N. K. Seymoar (2005). "The Liveable City." Vancouver Working Group Discussion Papers for the World Urban Forum. van den Bergh, J. and H. Verbruggen (1999). "An evaluation of the 'ecological footprint': reply to Wackernagel and Ferguson." Ecological Economics 31(3): 319-‐321. van den Bergh, J. C. J. M. (1996). Ecological Economics and Sustainable Development. Cheltenham, UK, Edward Elgar.
79 | P a g e
van den Bergh, J. C. J. M. and H. Verbruggen (1999). "Spatial sustainability, trade and indicators: an evaluation of the 'ecological footprint'." Ecological Economics 29(1): 61-‐72. VandeWeghe, J., R. and C. Kennedy (2007). "A Spatial Analysis of Residential Greenhouse Gas Emissions in the Toronto Census Metropolitan Area." Journal of Industrial Ecology 11(2): 133-‐144. Wackernagel, M. (1998). "The ecological footprint of Santiago de Chile." Local Environment 3(1): 7-‐25. Wackernagel, M., J. Kitzes, D. Moran, S. Goldfinger and M. Thomas (2006). "The Ecological Footprint of cities and regions: comparing resource availability with resource demand." Environment and Urbanization 18(1): 103-‐112. Wackernagel, M. and W. Rees (1995). Our Ecological Footprint: Reducing Human Impact on the Earth. Philadelphia, PA, USA, New Society Publishers. Warren-‐Rhodes, K. and A. Koenig (2001). "Ecosystem appropriation by Hong Kong and its implications for sustainable development." Ecological Economics 39(3): 347-‐359. WBCSD and WRI (2004). The Greenhouse Gas Protocol -‐ A Corporate Accounting and Reporting Standard, World Business Council for Sustainable Development and World Resources Institute. Weisz, H. and J. K. Steinberger (2010). "Reducing energy and material flows in cities." Current Opinion in Environmental Sustainability 2(3): 185-‐192. Wiedmann, T., M. Lenzen and R. Wood (2008). Uncertainty Analysis of the UK-‐MRIO Model -‐ Results from a Monte-‐Carlo Analysis of the UK Multi-‐Region Input-‐Output Model; Report to the UK Department for Environment, Food and Rural Affairs by Stockholm Environment Institute at the University of York and Centre for Integrated Sustainability Analysis at the University of Sydney. London, DEFRA. Wiedmann, T., J. Minx, J. Barrett and M. Wackernagel (2006). "Allocating ecological footprints to final consumption categories with input-‐output analysis." Ecological Economics 56(1): 28-‐48. Wiedmann, T., R. Wood, J. C. Minx, M. Lenzen, D. Guan and R. Harris (2010). "A Carbon Footprint Time Series of the UK -‐ Results from a Multi-‐Region Input-‐Output Model." Economic Systems Research 22(1): 19-‐42. Wolman, A. (1965). "The Metabolism of Cities." Scientific American: 179-‐190.
80 | P a g e
9 Annex A: Short literature review
9.1 Material flow analysis The literature concerned with the concept of urban metabolism and its quantification is filled with studies using Material Flow Analysis (MFA). MFA examines the material flows into a given system (private household, company, region, city, etc.), the material accumulations, so-‐flows within this system and the resulting outputs of the system. MFA can be applied to examine the relationship between a region or city and its corresponding hinterland (Hendriks et al. 2000). We distinguish material flow analysis from substance flow analysis (SFA), which is concerned with the analysis of individual or groups of chemical elements such as iron or carbon or chemical compounds such as carbon dioxide or iron chloride (Brunner and Rechberger 2004).
A simple search for the occurrence of urban metabolism in abstracts, keywords or titles of peer reviewed academic articles using the scientific full-‐text database www.sciencedirect.com20 resulted in 15 hits, of which the large majority were MFA studies. A wider search resulted in 95 articles referring to a wider mix of methodological approaches including SFAs, GIS based studies or studies focussing on the carbon cycle of cities.
Since the late 1970s a variety of cities and regions have been analyzed across the world including Brussels, Hong Kong, Sydney, London, Taipei, Beijing, Vienna, Swiss Lowlands, Lisbon or the Greater Toronto Area (Duvigneaud 1977, Newcome et. al. 1978, Warren-‐Rhodes and Koenig 2001, Newman 1999, Chartered Institute Of Wastes Management 2002, Huang and Hsu 2001, Huang and Chen (2009), Zhang et al. 2009, Baccini 1997, Obernosterer et al. 1998, Hendriks et al. 2000, Niza et al. 2009, Sahely et. al. 2003). Available studies typically focus on a single city (or region) at a single point in time. The main reasons for this are related to data availability and the work intensity of establishing material balances at the city level. Particularly for cross-‐sectional studies compiling a consistent pool of data is often difficult, when standard MFA approaches are used. The spatial boundaries for the analysis are most frequently defined by the administrative borders of the city itself or the larger metropolitan area.21 Further spatial granularity is usually not available.
The available literature varies considerably in terms of the type and variety of metabolic flows considered. Some analyses aim at establishing a comprehensive material balance as for example in cases studies for Vienna (Hendriks et al. 2000), Hong Kong (Warren-‐Rhodes and Koenig 2001) Limerick City Region (Browne et al. 2009) or Lisbon (Niza et al. 2009), while in some cases authors focus on specific metabolic flows such as water flows (Jenerette et al. 2006; Hubacek et al. 2009), construction materials (Huang and Hsu 2003) or food consumption (Barles 2007). While water, energy, waste and air emissions are the metabolic flows most frequently considered in MFA studies (see Box 2), there is much less evidence on other material flows such as biomass, minerals or consumer products (Kennedy et al. 2007).
Comprehensive MFA based studies can be usefully applied for the identification of critical (developments in) material flows and feed into local policy processes aimed at the re-‐design of the urban metabolism of a specific city in a preventive policy approach (Brunner 2007). It is the strength
20 www.sciencedirect.com currently contains 2500 peer-‐reviewed academic journals and more than 11000 books. 21 Regions are generally defined in terms of administrative boundaries.
81 | P a g e
of such a comprehensive MFA approach thmetabolism can be obtained. For example, cities might have implemented large-‐scale recycling and seen reductions in residential waste disposal, while other waste streams may still be on the rise. Equally, emission reductions in SO2 and particulates may have decreased, whereas NOX emissions might grow at the same time. Therefore, the comprehensiveness of such an accounting approach makes sure that shifts in environmental pressures between media can be detected and avoided.
The available evidence is less suited for comparative studies aiming at the generalisation of results due to the differences in data foundations and research methodology applied (Warren-‐Rhodes and Koenig 2001; Kennedy et al. 2007). Even though there are some common findings across studies such as the linear (rather than cyclic) metabmetabolism over time or the dependence of cities on their Hinterland (as material imports outweigh their exports by far), differences in the metabolism of cities and changes over time are usually quite varied. Partially this might be a true reflection of on-‐going trends, partially this might reveal differences in data foundations and methodologies (Kennedy et al. 2007).
A general limitation of the available studies are related to the extent indirect metabolic flows are considered. Studies mostly account only for the direct imports and exports and not the metabolic flows required further upstream to produce these products in a first instance (Niza et al. 2009).
9.2 Ecological Footprint Studies Despite the strong presence of MFA studies there are other approaches to measure urban metabolism. One concept, which has attracted attention in the last years, is the Ecological Footprint (EF). EF is defined as the total area of productive land and water required to continuously produce all the resources and assimilate all the wastes produced by consumption activities of a defined population. Environmental pressures arising from this consumption are captured in a composite land index.
Ecological Footprints have proven popular (Rees and Wackernagel 1996; Barrett 1998; Wackernagel 1998; Barrett and Scott 2001; Warren-‐Rhodes and Koenig 2001; Barrett et al. 2005; SEI et al. 2006; Wackernagel et al. 2006; Carballo Penela and Sebastián Villasante 2008; Owen et al. 2008; SEI 2008; Scotti et al. 2009). This might be partially due to the fact that EF can illustrate how cities depend on their Hinterland. As such the concept is an excellent communication tool (van den Bergh 1996).
Methodologically MFA studies have to be considered a particular type of MFA. However, once a material balance is established the flows are converted into land-‐units and become part of the energy footprint. The advantage of this is that compared to standard MFA the accounting of the indirect environmental pressures generated higher up in the global supply chain is more comprehensive. However, this comprehensiveness is only achieved by focussing on the energy content (and associated CO2 emissions) of the goods and services consumed in a city using conversion factors derived from life cycle inventory databases and other sources (Moran et al. 2007). Other indirect metabolic flows are not considered comprehensively.
82 | P a g e
Results typically show that cities require bio-‐productive land far greater than their own extent. In an what is
considered in the EF literature as their fair, sustainable (per capita) share (Wackernagel and Rees 1995; Barrett 1998; Barrett and Scott 2001; SEI et al. 2006; Carballo Penela and Sebastián Villasante 2008; Owen et al. 2008; SEI 2008; Scotti et al. 2009). At the same time the difference in estimates across regions and cities can be considerable (McDonald and Patterson 2003; McDonald and Patterson 2004; SEI et al. 2006).
Rees and Wackernagel (Rees and Wackernagel 1996) hypothesise that while the total EF of cities might be immense, their per capita footprint might be much lower than semi-‐urban or rural areas due to higher density and associated material and energy savings. Following this argument, cities as such might never be sustainable, but represent a key to sustainability. However, there is still relatively little evidence available. While UK evidence would support this claim at least for the South East of England Hu et al. (2008) show that this does not hold for young, rapidly developing cities in developing countries. A similar result is found by Hubacek et al. (2009)
However, a large part of the EF literature suffers from similar problems than MFA. Data availabilities at the local level is often poor and the construction of material balances (and ultimately EF accounts) time consuming. Given the differences in estimation methodologies and data foundations, results across studies are not always well suited for comparisons. A quantitative comparison of results from a variety of different EF studies can be found in Scotti et al. (2009).
Rather recently authors have therefore started estimating EFs in generalised input-‐output models (Bicknell et al. 1998; McDonald and Patterson 2004; Wiedmann et al. 2006; Carballo Penela and Sebastián Villasante 2008). These models allow a consistent assignment of EFs to final consumption activities and therefore more comparable studies. Using such an approach SEI (2006) provides Ecological Footprint estimates for all 411 municipalities in the UK, while McDonald and Patterson (2004) quantifies the EF of 16 New Zealand region. The two our knowledge only spatially disaggregated studies using a conventional EF approach is provided by Bagliani et al. looking at 36 municipalities in the Sienna region in Italy (Bagliani et al. 2008). Muniz and Galindo (2005) provide a particularly interesting approach. Quantifying the EF of commuting in 153 municipalities in the Barcelona region they find urban form to be the main determinant of EF variations.
However, EF is a composite indicator of climate change and land use combining real with a notion of hypothetical land (van den Bergh 1996; van den Bergh and Verbruggen 1999; van den Bergh and Verbruggen 1999). The hypothetical land is converted energy. Such a conversion is always arbitrary. Even though the EF is perfectly suited for environmental communication purposes, such a composite indicator is not adequate for informing policy.
9.3 Other studies There are a variety of other studies in the wider field of urban metabolism focussing on specific metabolic flows. The largest and most rapidly growing share of literature focus on energy and CO2 associated with cities and regions even though there is also evidence for water, land use or local air emissions (Jenerette et al. 2006; Schwela et al. 2007; Lenzen and Peters 2009).
The large number of studies on energy use and CO2 emissions in cities might be best distinguished according to the type of inventory (production based or consumption based), the
83 | P a g e
comprehensiveness of the account as well as the spatial detail of the study. There are a variety of production based studies providing energy, CO2 or GHG estimates for different cities. Dhakal (2009) compares energy use and CO2 emissions of the 35 largest cities in China using a top-‐down approach based on regional GDP and emission data for the time period 1995-‐2006. The urban areas of larger city provinces are distinguished based on administrative boundaries at the county-‐level. Using index decomposition analysis the drivers of changes in CO2 emissions are analysed. Similar emission studies are provided elsewhere (ICLEI and Carbon Disclosure Project 2008; Dodman 2009).
However, with on-‐going standardisation processes in producing local city-‐scale emission inventories there has been an increasing acknowledgment that inclusion consumption based emissions (scope 2 and 3 emissions) in studies as well. Most of the available studies only provide consumption based estimates in specific areas like air transportation, sewage treatment etc. (Kennedy et al.; Ramaswami et al. 2008; Kennedy et al. 2009). Browne et al. (2008) for example develops a methodology for estimating a partial carbon footprint for the 100 largest metropolitean areas in the U.S. covering electricity consumption, heating fuels as well as highway transportation. However, the growing mix of estimation methodologies also comes with problems. Dodman (2009), for example, highlights the difficulties of comparing emission estimates across studies and regions.
Comprehensive assessments of energy use and consumer emissions are provided only in comparatively few studies (Lenzen et al. 2004; Druckman et al. 2008; Lenzen and Peters 2009; Minx et al. 2009). These studies are all based on generalised input-‐output models. An advantage of this approach is that a set of estimates for different regions and cities can be derived from an consistent estimation framework.
Interestingly, these studies also provide spatially granular estimates. Minx et al. (2009) provide middle-‐layer super-‐output area (<10000 households) carbon footprint estimates for the UK. Such spatial disaggregations often reveal considerable differences in the carbon footprint of cities between rural and urban portions of metropolitean areas (particularly at the edges) and give a much better impression of environmental pressures of urbanisation processes and sprawl. Furthermore, these estimates are based on detailed lifestyle data and are well suited to analyse drivers behind CO2 emissions (Baiocchi et al. 2009). A similar approach and similar levels of granularity is achieved by Druckman et al. (2008) for energy use.
Other studies providing spatially granular estimates are either based on producer emission estimates (AEA Technology 2008) or focus on energy consumption (mainly private motoring, housing) (VandeWeghe and Kennedy 2007; Brown et al. 2008; Parshall et al. 2009). While the revealed trends in GHG emissions from urban development is confirmed, methodologically these approaches are of interest as they are much less data intensive than the input-‐output based ones.
The only study that currently manages to bridge the gap between a production and a consumption based inventory on a sub-‐national level is provided by Lenzen and Peters (2009). Linking gridded information on economic activity and emission sources for 1400 Statistical Local Areas into a generalised input-‐output model for Australia, the authors can show how consumption activities in the cities of Melbourne and Sydney trigger emissions (or water or employment) in different parts of country. This is particularly valuable when local pollutant such as water are concerned, where the environmental impacts generated heavily depends on where the water comes from (i.e. water rich or water scarce area).