Assessing the Ecological Footprint of a Large Metropolitan Water ...

Journal of Environmental Planning and Management, 46(1), 113–141, 2003

Assessing the Ecological Footprint of a Large Metropolitan Water Supplier: Lessons for Water Management and Planning towards Sustainability

MANFRED LENZEN*, SVEN LUNDIE†, GRANT BRANSGROVE‡, LISA CHARET‡ & FABIAN SACK‡ *School of Physics, A28, University of Sydney, Sydney, NSW 2006, Australia. E-mail: [email protected] † Centre for Water and Waste Technology, H22, Vallentine Annex, University of New South Wales, Kensington 2033, Australia ‡ Sydney Water Corporation, 115–123 Bathurst St, Sydney, NSW 2000, Australia (Received February 2002; revised September 2002)

ABSTRACT Faced with the task of communicating their combined social, environmental and economic impact, water service providers are seeking to report overall performance in an aggregated way. Such a methodology must be scientifically robust, easily communicated and allow benchmarking of performance while reflecting a transition towards sustainability. In this paper the ecological footprint (EF) is calculated for Sydney Water Corporation (SWC), using input–output analysis and land disturbance in an innovative approach that overcomes problems identified in the original EF concept. This pilot study has allowed SWC to gain some valuable insights into its impacts: SWC’s annual EF is about 73 100 ha in terms of land disturbance. Of this, 54 000 ha are projected to become disturbed as a consequence of climate change, with the remainder of 19 100 ha being disturbed on SWC’s premises (2400 ha) and on those of upstream suppliers (16 700 ha). Total on-site impacts equal 9300 ha, while indirect land disturbance contributes 63 600 ha. The EF appears promising as an educational and communication tool and may have potential as a decision support tool. However, further research is needed to incorporate downstream impacts into the EF, which would have significant benefits to SWC in terms of assessing and communicating the organization’s overall progress towards sustainability.

Introduction The carrying capacity of an environment is defined as its maximum persistently supportable load, and it can be considered as “the maximum population of a given species that can be supported indefinitely in a defined habit without permanently impairing the productivity of that habitat” (Wackernagel & Rees, 1995, p. 224). Because of this dependency, the ecological footprint (EF) concept has been developed by Rees and Wackernagel as a tool that enables the determination of the area of land (and water) in various categories that is 0964-0568 Print/1360-0559 Online/03/010113-29  2003 University of Newcastle upon Tyne DOI: 10.1080/0964056032000037168

114 M. Lenzen, S. Lundie, G. Bransgrove, L. Charet & F. Sack required on a continuous basis to provide all energy and material resources consumed and to absorb all waste discharged by a single person, a specific population, a specific economy or the entire world (Rees, 1992; Wackernagel & Rees, 1995; Rees, 1996; Wackernagel et al., 1999; Rees, 2001). The EF concept has been further developed and has gained more significance since its first introduction. However, the original focus of EF case studies has been on geographical entities such as countries (52 countries (Wackernagel et al., 1997), Australia (Simpson et al., 2000; Lenzen & Murray, 2001), New Zealand (Bicknell et al., 1998), Benin, Bhutan, Costa Rica and the Netherlands (van Vuuren & Smeets, 2000)), regions (south-east Queensland (Simpson et al., 1998), Guernsey (Barrett, 2001), Isle of Wight (Best Foot Forward & Imperial College, 2001)) and cities (Santiago de Chile (Wackernagel, 1998), Vancouver City and the Lower Fraser Basin (Rees & Wackernagel, 1996), Canberra (Close & Foran, 1998)). More recently, EFs have been evaluated along functional lines, including investigations of universities (Newcastle (Flint, 2001), Redlands (Venetoulis, 2001), Sydney (Wood & Lenzen, 2002)), as well as products and packaging systems (Lewis et al., 2000). There have only been a limited number of cases where the EF has been used to indicate the environmental sustainability of complete commercial organizations (for example Anglian Water Group, 2001; Chambers & Lewis, 2001). In this paper an EF calculation for Sydney Water Corporation (SWC) is presented. The motivation of SWC for this innovative project is outlined in the next section. A detailed description of EF methodology is given in the following section, followed by SWC’s EF results, their discussion and an analysis of their planning and management relevance, before the paper is concluded.

SWC’s Motivation to Apply the EF Concept SWC is Australia’s largest water and wastewater service provider. SWC provides water, wastewater and some stormwater services to the greater Sydney area, distributing and retailing bulk water supplied by the Sydney Catchment Authority. SWC operates as a statutory state-owned corporation under the Sydney Water Act 1994. This Act enables the provision of water services to a community of approximately 4 million, emphasizing the importance of protecting public health, equitable access to its services and contributing to the state’s economy. The Act sets as a principal objective to protect the environment by conducting operations in compliance with the four principles of ecologically sustainable development (ESD) established by the 1992 Australian Inter-governmental Agreement on the Environment. SWC has interpreted these as follows. • Precautionary principle. Reduce the chance of serious or long-term environmental problems, even if we are not sure that these problems will occur. • Inter-generational and intra-generational equity. Reduce the effects of activities on the environment that the community, now and in the future, relies on to meet its needs and expectations. • Conservation of biological diversity and ecological integrity. Maintain or enhance the range of native plants and animals and the health of natural areas.

Assessing the Ecological Footprint of a Large Metropolitan Water Supplier

115

• Improved valuation and pricing of environmental resources. Improve the way that we undertake valuation of environmental costs and benefits and use this information when making decisions. SWC is committed to the Australian goal for ESD: “development that improves the quality of life, both now and in the future, in a way that maintains the ecological processes on which life depends” (Commonwealth of Australia, 1992, p. 8). SWC’s policy is to implement the principles of ESD by developing long-term strategies and plans that are informed by community consultation. SWC has sought to adopt tools that assist in identifying overall cumulative environmental and social impacts and is moving towards integrated ecological, social and financial reporting (triple-bottom-line reporting). SWC recognizes that developing indicators and measuring sustainability is a rapidly developing field and is undertaking a range of activities aimed to achieve continuous improvement. In particular SWC’s ESD indicators give greater emphasis to environmental impacts over social and economic impacts. This can partially be attributed to the nature of the corporation’s core business, but is also the result of a historical emphasis on the ecological basis of sustainability within the Australian institutional context. Data collection against 29 selected ESD indicators (see Glynn & Sack, 2001; SWC, 2001a), commenced in July 2000, and public reporting commenced in October 2001 with the publication of SWC (2001b). Performance against any one of the ESD indicators only provides data in one sustainability parameter and does not truly indicate sustainability. For example, a reduction in pollutants or discharges to the environment may result in an increase in energy consumption. If the results for all considerations in this process are not measured or reported, the integration of social and environmental benefits cannot be shown to be balanced against the expenditure required to achieve this, or possible decreases in other environmental factors. Faced with the task of communicating the combined social, environmental and economic impact measured by the indicators, SWC sought a methodology that would allow at least some measure to be cumulatively reported. The challenges faced by any methodology aimed at cumulating indicators of sustainability are considerable. The methodology must: • be sufficiently robust to attract wide credibility (especially from technical and scientific audiences); • provide a means for stakeholders and the community to understand what the performance against ESD indicators means in terms of cumulative impacts and progress towards ESD; • be transferable (at least in principle) to allow benchmarking with other industry leaders. Perhaps most importantly, the methodology must provide a metric into which other incommensurable sustainability indicators can be translated and cumulated. For example, SWC’s ESD indicators include measures of greenhouse gas emissions (tonnes of carbon dioxide equivalent), waste generation (tonnes of waste generated, disposed to landfill and recycled) and user pays (percentage of total water and sewerage revenue based on usage charges) (see SWC, 2001b). Scanning of international best practice revealed a case study of a promising

116 M. Lenzen, S. Lundie, G. Bransgrove, L. Charet & F. Sack methodology: EF (as described in the ‘Methodology’ section). Hence, a pilot EF calculation was undertaken in order to establish to what extent it could meet the challenges of reporting combined progress towards sustainability. The results of the pilot SWC EF were reported in SWC (2001b). A range of methodological issues have become apparent, including: the scope of sustainability measures capable of being included within the footprint; the robustness of the methodology; and the relevance of the methodology to organizational process. Methodology SWC’s EF has been assessed, based on four existing methods relevant in the Australian context and proposed by: (1) Rees (1992), Wackernagel & Rees (1995), Simmons et al. (2000), Barrett (2001) and Chambers & Lewis (2001); (2) Simpson et al. (2000); (3) Bicknell et al. (1998); and (4) Lenzen & Murray (2001). The main features of these methods are described below, followed by a more detailed explanation of the approach taken for SWC. Development of the EF Concept The EF was originally conceived as an intuitively simple and elegant method for comparing the sustainability of resource use among populations (Rees, 1992). The consumption of these populations is converted into a single index: the total area of land and water ecosystems required on a continuous basis to produce their resources and assimilate their wastes, wherever these areas are located. In order to measure the sustainability of a given population or country, the EF is compared to the available land area. Unsustainable populations, in this analysis, are populations with a larger EF than their domestic land base and no possibility of importing carrying capacity through trade (see Rees, 2001). Since its initial formulation, the EF has been criticized by a number of researchers (Levett, 1998; van den Bergh & Verbruggen, 1999; Ayres, 2000; Moffatt, 2000; Opschoor, 2000; Rapport, 2000; van Kooten & Bulte, 2000; Lenzen & Murray, 2001; Ferng, 2002). The objections largely refer to the oversimplification of the complex task of measuring the sustainability of consumption. In particular, perceived shortcomings exist in the method’s impact weighting, sequestration scenario and spatial delineation (Levett, 1998; van den Bergh and Verbruggen, 1999; Ayres, 2000; Lenzen & Murray, 2001). Most of the critics believe that EFs are inadequate for (regional) policy design (van den Bergh & Verbruggen, 1999; Ayres, 2000; Moffatt, 2000; Opschoor, 2000; van Kooten & Bulte, 2000; Lenzen & Murray, 2001), and that they claim but do not enable an analysis of unsustainability (van den Bergh & Verbruggen, 1999; Rapport, 2000; Lenzen & Murray, 2001). The concept is generally acknowledged as a valuable educational tool to highlight global (not regional) unsustainability, and has helped to reopen the debate on human carrying capacity in the context of global sustainability. It has limited use as a policy and planning tool for indicating unsustainability, because it does not reveal where impacts really occur, what the nature and severity of these impacts are and how these impacts compare with the self-repair capability of the respective ecosystem (the temporal aspect, see Arrow et al., 1995). In response to these problems, significant modifications have been proposed (Bicknell et al., 1998; Lenzen & Murray, 2001; Ferng, 2002).


117

The original EF method refers to land which is ecologically productive for human purposes (Wackernagel & Rees, 1995, p. 11). However, even unproductive land can be useful for human purposes, and should not be disregarded. This has been acknowledged by Simpson et al. (2000), who include semi-arid and arid land in their assessment of Australia’s EF. Therefore, the calculations in this work consider all areas of land. In the conventional EF, land needed for producing commodities consumed by populations is assessed in a bottom-up approach, covering mostly land used directly by the respective producer. Land used by suppliers of producers is often not taken into account. In other words, the analysis is only one or two production layers deep.1 While this approach is unproblematic at the national level, it leads to misallocation and systematic truncation errors at the component, institutional and regional level (Lenzen & Murray, 2001). The first authors to propose the use of input–output analysis for calculating comprehensive EFs were Bicknell et al. (1998), who assessed the EF of the New Zealand population. Because of its intrinsic features, the input–output approach guarantees the complete coverage of upstream land and emission requirements up to an infinite order (see the section below, ‘Input–output analysis’). Simpson et al. (2000) also take advantage of input–output analysis, but only by using input–output-based intensities for carbon dioxide, and not for land. In conventional EFs, land categories are weighted with equivalence and local yield factors (Wackernagel et al., 2002a) in order to express appropriated bioproductivity in world-average terms. However, the intensity of human-induced changes to land is independent of productivity (compare van den Bergh & Verbruggen, 1999). In the use of appropriated world-average productivity, the original approach does not yield insights about regional impacts on land and ecosystem processes, and therefore does not indicate the sustainability of regional land use (Rapport, 2000; Lenzen & Murray, 2001). Lenzen & Murray (2001) propose that the EF should be primarily concerned with land disturbance, and that this approach would better reflect the image of a ‘footprint on land’, because it describes the effects of human land use on ecosystems, independent of productivity. They concede, however, that a disturbance-based approach still cannot address whether land use is practised sustainably. Nevertheless, land cover disturbance is a contributor to unsustainability of farming or forestry practices, as it can be a precursor to soil erosion and other causes of land degradation (Graetz et al., 1995). Until more detailed data are available on the unsustainability of various activities, land cover disturbance is the best indication available (compare van den Bergh & Verbruggen, 1999). For this reason, the present authors use modified Australian land types and a weighting system (land conditions, C in Table 1) for these types, which reflect the degree of alteration of land from its natural state. ‘Energy land’ is calculated conventionally using either a ‘carbon sequestration’ factor (Wackernagel & Rees, 1995) or a ‘fuelwood equivalence’ factor (Wackernagel et al., 2002a). Hypothetical ‘fuelwood’ land is also responsible for the global ‘overshoot’ (Wackernagel et al., 2002b) or the ‘carbon sink deficit’ (Rees, personal communication). Rather than this bioproductivity measure for compensating emissions, and consistent with the disturbance-based land use approach, the present authors consider the projected disturbance of terrestrial and aquatic ecosystems due to climate change and sea level rise for the incorporation of energy use and greenhouse gas emissions (including non-carbon dioxide gases

118 M. Lenzen, S. Lundie, G. Bransgrove, L. Charet & F. Sack Table 1. Weights C describing the alteration from the natural state for land types (derived from Hobbs & Hopkins, 1990; Graetz et al., 1995; Swan & Pettersson, 1998; van Dobben et al., 1998; Ko¨llner, 2000; see also Lindeijer, 2000), area A affected in Australia and resulting land disturbance D ⫽ A ⫻ C (in units of 106 ha) Land typec

Affected area (A)

Land condition (C)

Consumed Built

2.3 2.3

1.0

2.3 2.3

16.4 15.5 0.8 0.2

0.8

13.2 12.4 0.6 0.1

Replaced Crop land Cleared, ILZ Non-native coniferous planatations

101.1 15.5 84.7 0.9

0.6

60.7 9.3 50.8 0.5

Disturbed Thinned, ILZ Significantly disturbed, ELZ Reversibly built Native eucalypt plantations

162.8 47.1 115.2 0.3 0.2

0.4

65.2 18.8 46.1 0.2 0.1

Partially Disturbed Indeterminately disturbed, ILZa Substantially disturbed, ELZb

102.1 36.5 65.6

0.2

20.4 7.3 13.1

Slightly disturbed Uncleared, ILZ Slightly disturbed, ELZ Reserves and unused Crown land

378.1 85.9 8.3 284.0

0.0

0.0 0.0 0.0 0.0

Total

768.2

Degraded Degraded pasture Degraded crop land Mined land

Disturbance (D)

161.8

a

Areas for which disturbance could not be assessed. Disturbance below-critical for biotic erosion. c ELZ ⫽ Extensive Land Use Zone (Central, Western and Northern Australia). ILZ ⫽ Intensive Land Use Zone (Eastern and South East Australia). b

and non-energy sources) into the EF. For doubled carbon dioxide equilibrium conditions, climate change scenarios (Intergovernmental Panel on Climate Change, 1995; Darwin et al., 1996) yield a conversion factor of 68.5 ha/kt CO2–e/a (see Lenzen & Murray, 2001)2. Due to the substantial uncertainty in predicting impacts of climate change on land, these figures should be taken as a crude approximation. Combining present and potential future land disturbance, the authors define the EF as a sum of areas Ai of i ⫽ 1, … ,7 land types (six in Table 1 plus ‘emissions land’) weighted with land condition factors Ci: EF ⫽ i Ai Ci

(1)

Finally, since the original approach was not designed for institutional EFs, the authors follow Barrett (2001) and Simmons et al. (2000) and use local (not world-average) land and emission figures in a detailed commodity (‘component’) approach. The three fundamental approaches to calculating the EF are illustrated in Figure 1: appropriated bioproductivity (approach 1.1: Rees (1992), Wackernagel & Rees (1995), Wackernagel et al. (1997, 2002a), Simmons et al. (2000), Barrett (2001), Chambers & Lewis (2001)),3 including arid lands and input–output-based


119

Figure 1. Schematic diagram visualizing the relationship between different EF methods relevant in the Australian context.

carbon dioxide intensities (approach 1.2: Simpson et al. (2000)); input–output analysis and land use (approach 2: Bicknell et al. (1998)); and input–output analysis and land disturbance (approach 3: Lenzen & Murray, 2001). Input–Output Analysis In this study, a hybrid EF approach is employed, combining input–output analysis with process analysis. In this approach, the direct (on-site) land and emission requirements of SWC are assessed in a detailed company audit, while all remaining higher-order requirements (for materials extraction, manufacturing and services) are covered by input–output analysis. An example of a higher-order requirement is the land mined by iron ore mining operators to provide the iron ore for the steel to manufacture construction machinery used in SWC’s infrastructure works. Hybrid techniques have been applied previously in a number of life-cycle assessments (compare Bullard et al., 1978; Lave et al., 1995; Treloar, 1997; Hondo & Sakai, 2000; Joshi, 2001; Lenzen, 2001a; Suh & Huppes, forthcoming). Input–output analysis is a top-down economic technique which uses sectoral monetary transaction data to account for the complex interdependencies of industries in modern economies. Generalized input–output frameworks have been applied extensively to environmental analysis since the late 1960s (see, for example, Isard et al., 1967; Leontief & Ford, 1970). The result of generalized input–output analyses is a f ⫻ n matrix of factor multipliers, that is, embodiments of f production factors (here: land types and greenhouse gas emissions) per unit of final consumption of commodities produced by n industry sectors. A multiplier matrix M can be calculated from an f ⫻ n matrix F containing sectoral production factor usage, and from a n ⫻ n direct requirements matrix A according to: M ⫽ F(I ⫺ A) ⫺ 1

(2)

120 M. Lenzen, S. Lundie, G. Bransgrove, L. Charet & F. Sack where I is the n ⫻ n unity matrix. The f ⫻ 1 land and emissions inventory of SWC (f ⫽ 7, 6 land types and greenhouse gas emissions) is then determined by multiplying SWC’s annual operation cost (represented by an n ⫻ 1 commodity input vector y) with the multiplier matrix M, and adding a f ⫻ 1 vector d of direct (on-site) land and emissions: ⫽ M ⫻ y ⫹ d

(3)

M ⫻ y represents f ⫽ 7 indirect requirements, that is, land and emissions embodied in all inputs into SWC’s operation. From the 7 elements of , land disturbance and the EF are calculated residually. An introduction to the input–output method and its application to environmental problems can be found in papers by Leontief & Ford (1970), Duchin (1992) and Dixon (1996). The mathematical formalism used to derive equations (2) and (3), and some of the results presented in this paper, are described in detail in a previous paper (Lenzen, 2001b). Errors in Input–Output-based EFs While being able to cover an infinite number of production stages in an elegant way, input–output analysis suffers from uncertainties arising from a number of sources such as source data sampling and reporting errors, assumptions about foreign industries, the assumption of proportionality between monetary and physical flow, the aggregation of input–output data over different producers and the aggregation of input–output data over different products supplied by one industry (Lenzen, 2001a). Standard errors Mij of elements in the multiplier matrix M defy analytical treatment, and can therefore only be determined using stochastic analysis (Bullard & Sebald, 1977, 1988). The Mij as used in this work were calculated by Monte Carlo simulations of the propagation of numerical perturbations from F and A through to M. Given the standard errors Mij, and the standard errors yj and d,i of y and d, the total standard error i of an element in the factor inventory vector is:

冑冘 (y M ) ⫹ 冘 (M y ) ⫹ n

i ⫽

k⫽1

n

k

ik

2

k⫽1

ik

k

2

2 dj

(4)

The standard errors Mij, yj and d,i are assumed to be stochastic (Quandt, 1959). This feature enables the total standard error i to decrease with increasing number of non-zero entries in y, that is, with increasing detail of the breakdown of the inputs for SWC’s annual operation. This can be seen as follows. Let d,i«, and yj, j ⫽ 1, … ,m be m non-zero entries in y, so that for factor i:

冘M y m

i ⬇

k⫽1

ik k

(5)

Let all yk, Mik, yk and Mik be approximately of the same order, that is: yk ⬇ y, Mik ⬇ Mi, yk ⬇ y, and Mik ⬇ Mi ∀k ⫽ 1, … m. The relative standard error i/i of i is then:

(6)


冑冘 (y M ⫹ 冘 (M y ) m(y M ) ⫹ m(y M ) ⬇ ⬇冑 ¯¯ ) (m y¯ M 冘M y M) y 1 冑 ⫽ ⫹冉冊冉冊 ¯¯ y¯ 兹m M m

m

k

k⫽1

i

121

ik

ik

k⫽1

k

2

2

i

i

m

i

i

2

ik k

k⫽1 i

2

2

2

(7)

i

In order to minimize the relative standard error of the land and emissions inventory, it is therefore important to (1) obtain a breakdown of the inputs of SWC’s operational cost that is as detailed as possible (large m) and (2) obtain important direct land requirements and emissions with low relative standard errors d,i/d,i. In the case of SWC, the account breakdown contains m ⫽ 42 entries in the input–output product classification (IOPC), which is sufficient to guarantee small errors of indirect land and emission requirements. The errors yi and d,i of financial data yi, and of on-site land use and greenhouse gas emissions d,i, were assumed to be zero, since these quantities are well known by the SWC. Under these assumptions, even for a conservative estimate of relative errors of multipliers, say Mi/Mi ⬇ 100%, the error of the EF of SWC can be estimated to be (EF) ⬇ 100%/兹42 ⬇ 15%. Structural Path Analysis The general decomposition approach described in the following section was introduced into economics and regional science in 1984 under the name structural path analysis (Crama et al., 1984; Defourny & Thorbecke, 1984), and applied in life-cycle assessment by Treloar and Lenzen (Treloar, 1997, 1998; Treloar et al., 2000; Lenzen, 2001c). The total factor multipliers as in equation (2) can be decomposed into contributions from structural paths, by ‘unravelling’ the Leontief inverse using its series expansion: F (I ⫺ A) ⫺ 1 ⫽ F ⫹ FA ⫹ FA2 ⫹ FA3 ⫹ …

(8)

Expanding equation (8), indirect land and emission requirements Mi ⫻ yi as in equation (3) can be written as:

冘 F ( ⫹ A ⫹ (A ) ⫹ (A ) ⫹ …) ⫽ y 冘 F ( ⫹ A ⫹ 冘 A A ⫹ 冘冘 A A A …) ⫽Fy ⫹ 冘 FA y ⫹ 冘 F 冘 A A y ⫹ 冘 F 冘 A 冘 A A y ⫹ … n

Miyi ⫽ yi

j⫽1 n

i j⫽1

j

ij

ji

j

ji

ji

3

ji

ji

n

n

i i

2

j⫽1

j

ji i

j⫽1 n

k⫽1

n

jk

ki

n

l⫽1 k⫽1

n

k

j⫽1

jl

lk

n

kj

ji i

l⫽1

ki

n

l

k⫽1

n

lk j⫽1

kj

ji i

(9)

where i, j, k and l denote industries, and ij ⫽ 1 if i ⫽ j and ij ⫽ 0 otherwise. Miyi is thus a sum over a direct factor input Fiyi, occurring in industry i itself, and higher-order input paths. An input path from industry j (domestic or foreign) into industry i of first order is represented by a product FjAjiyi, while an input path from industry k via industry j into industry i is represented by a product FkAkjAjiyi, and so on. There are n input paths of first order, n2 paths of second order and, in general, nN paths of Nth order. An index pair (ij) shall be referred to as a vertex.

122 M. Lenzen, S. Lundie, G. Bransgrove, L. Charet & F. Sack Equation (9) was evaluated by sequential backwards scanning of the production chain tree from final demand to the various locations of production factor usage. The result of one execution of this algorithm for a particular production factor is a ranking of input paths for each of the n (135 in this work) sectors in terms of their contribution to the total EF.

Results The calculation of SWC’s EF measures proceeded in three steps. First, emissions as well as land use and disturbance multipliers (M, see equation (2)) were determined using the 1994–95 Australian input–output tables (Australian Bureau of Statistics, 1999a), the National Greenhouse Gas Inventory (National Greenhouse Gas Inventory Committee, 1998) and land use and condition data from various sources (see Lenzen & Murray, 2001). Secondly, SWC’s annual account for the financial year 2001 was reclassified and compressed into the IOPC (Australian Bureau of Statistics, 1999b), yielding a commodity input vector y. An indirect land and emissions inventory (M ⫻ y, see equation (3)) was then obtained by multiplying each item in the compressed account with its respective multiplier. SWC’s final EF was calculated residually using weighting and conversion factors corresponding to the different EF methods. SWC’s EF was further analysed using a decomposition technique and structural path analysis. While a decomposition yields a breakdown of land and emissions into contributions from upstream production layers (equation (8)), structural path analysis yields a ranking containing the most important structural paths (equation (9)). Both approaches provide underlying information illustrating the origins and causes of SWC’s EF. In the following, the results of all calculations will be presented in order of increasing detail, that is, aggregate results first, followed by a breakdown into land types and greenhouse gas emissions, a further breakdown into commodities supplied to SWC, a decomposition into production layers and, finally, structural paths.

SWC’s EF Table 2 presents a summary of SWC’s EF as calculated according to four methods. The table headers indicate whether the respective column refers to bioproductivity, land use or land disturbance, and which conversion factors for energy/emissions were used. The rows distinguish on-site impacts (land used and greenhouse gas emissions arising directly through SWC activities), first- and second-order industrial impacts (land and greenhouse gas requirements occurring directly on the premises of producers supplying commodities to SWC) and total industrial (infinite upstream) requirements. The sum of on-site and industrial requirements is called total requirements. The calculations according to Rees (1992), Wackernagel & Rees (1995), Simmons et al. (2000), Barrett (2001), Chambers & Lewis (2001) and Simpson et al. (2000) were assumed to cover all first- and second-order impacts (see Note 1). SWC’s EF, when assessed using the original concept, yields an area of 95 400 ha. Of this footprint, 8300 ha is directly caused on the premises of SWC, in the form of land occupied (water and sewage pumping stations, water and sewage treatment plants, trunk drainages, stormwater channels, sewers, reservoirs,


123

Table 2. Summary: SWC’s EFs according to various approaches (Kha)

Conversion factor On-site First- and secondorder All industrial Total d

Land Greenhouse gases

Bioproductivity (Rees, Wackernagel, Simmons, Barrett, Chambers, Lewis) 0.0163 gha/GJa,b

Bioproductivity (Simpson et al.) 0.0163 gha/GJa,b

Land use (Bicknell et al.) 0.01 ha/GJb

Land disturbance (Lenzen & Murray) 68.5 ha/kt

8.3

8.3

6.8

9.5

87.1

87.2 115.1c

124.3

63.6

95.4

123.3

131.2

73.1

7.5 87.9

7.6 115.7

49.5 81.7

19.1 54.0

Fuelwood method; gha ⫽ global hectares. Energy-related carbon dioxide emissions only. c Bioproductivity up to 2nd order, energy-related carbon dioxide emissions for all orders. d In terms of bioproductivity, land use of disturbance as described in header. This land use does not include slightly disturbed land (C ⫽ 0; mostly catchments and reserves, which were assumed to be in a near-pristine state). a

b

depots, computer centres and canals, etc.) and carbon dioxide emissions (from fuel use); 87 100 ha are required in the form of bioproductivity appropriated by suppliers of commodities for SWC. Of the total, only 8% is caused by land use, but 92% is needed to compensate carbon dioxide emissions. Once the full life-cycle of commodities is taken into account, the EF increases significantly: SWC’s EF, according to Simpson et al. (2000), is larger because higher-order carbon dioxide emissions are taken into account. Similar increases would occur when including higher-order upstream land requirements. In contrast, the inclusion of arid land only increases the footprint by 100 ha. This is because SWC’s on-site, first- and second-order inputs do not originate from agricultural industries, which predominantly occupy arid land, but from manufacturing, construction and services. Including all upstream requirements of land and carbon dioxide emissions (Bicknell et al.), SWC requires about 131 200 ha of land. An interesting fact is that the fraction of actual land in this footprint has increased from below 10% to about 35%. This is due to the fact that in a completed upstream scenario, agricultural industries provide some higher-order inputs into the supply chains leading to SWC. These inputs are associated with large land uses, and therefore contribute relatively more landed than greenhouse gas emissions. These higherorder effects are illustrated in the sections below, ‘Decomposition into upstream production layers’ and ‘Structural path analysis’. Weighting land impacts, SWC’s disturbance-based EF measures 73 100 ha; this corresponds to approximately 178 m2 per capita. Within this area, 54 000 ha are projected to become disturbed as a consequence of climate change, and 19 100 ha are disturbed through use on SWC’s premises (2400 ha) and its suppliers’ (16 700 ha). It must be emphasized that the conversion of greenhouse gas emissions into a land equivalent is associated with considerable uncertainty. Assuming a standard error of 100% for the conversion factor of 68.5 ha/kt CO2-e, SWC’s total disturbance-based EF should be reported as being between 45 and 120 kha.

124 M. Lenzen, S. Lundie, G. Bransgrove, L. Charet & F. Sack Breakdown into Land Types and Greenhouse Gas Emissions As a first step in decomposing the aggregate result shown in Table 2, SWC’s EF is broken down into the six land types, their composite land disturbance and greenhouse gas emissions (see Table 3). Rows are organized in the same way as the first four rows of Table 2. The on-site EF is, as mentioned before, dominated by greenhouse gas emissions (methane and carbon dioxide from fuel use). On-site land disturbance is 2400 ha, mainly in the form of consumed land (C ⫽ 1; water and sewage pumping stations, water and sewage treatment plants, trunk drainages, stormwater channels, sewers, reservoirs, depots, computer centres and canals, etc.). Total first- and second-order land requirements are small, since SWC does not require inputs from land-intensive industry sectors such as agriculture and forestry, which dominate most of the categories with C ⱕ 0.8, but mostly from manufacturing, construction and services, which exhibit relatively low land multipliers, mostly at C ⫽ 1. However, this situation changes dramatically when all upstream requirements are taken into consideration, and higher-order inputs from land-intensive agricultural industries become important: indirect land disturbance increases eightfold. This will be shown more quantitatively in sections below, ‘Decomposition into upstream production layers’ and ‘Structural path analysis’. The breakdown of greenhouse gas emissions yields similar results: third- and higher-order requirements are of the same magnitude as on-site to second-order requirements. On-site emissions are mainly in the form of methane and carbon dioxide from fuel combustion, while first-order emissions mainly originate from coal-fired power stations generating electricity for SWC (see sections below, ‘Breakdown into commodities supplied to SWC’ and ‘Structural path analysis’). A noticeable part of direct land use is contained in the category ‘slightly disturbed’, which comprises catchment areas owned by SWC (275 400 ha) and embodied in SWC’s bulk water purchases from the Sydney Catchment Authority (1 660 000 ha). The allocation of these catchment areas to the land category ‘slightly disturbed’ (C ⫽ 0) was carried out because of three reasons. First, there is a lack of information on the land condition of catchments, whether owned by SWC or managed by the Sydney Catchment Authority. Secondly, the land disturbance occurring across large areas of the catchment results from agriculture, mining and urban development, which have nothing to do with water supply catchment activities. Therefore, assigning all of this land to water supply would not give an accurate reflection of SWC’s impact on the environment. Thirdly, the inner catchment areas managed by the Sydney Catchment Authority purely for water supply are still in a near-pristine state. In fact, catchment protection ensures an environmental benefit to this land by prohibiting development and other ecological risks.

Breakdown into Commodities Supplied to SWC Table 4 shows the same information as in Table 3, but at the IOPC-classified commodity level. The purchase of electricity by SWC is the single most important item contributing to land disturbance and the EF, followed by contracted non-residential construction work. The category ‘water supply’, ranking third,

2.1 1.3 1.8

3.8

On-site First- and second-order All industrial

Total

Consumed land (C ⫽ 1) (kha)

2.0

0.2 0.1 1.8

Degraded land (C ⫽ 0.8) (kha)

9.3

0.0 0.4 9.3

Replaced land (C ⫽ 0.6) (kha)

15.5

0.2 0.9 15.3

Significantly disturbed land (C ⫽ 0.4) (kha)

9.3

0.1 0.4 9.2

Partially disturbed land (C ⫽ 0.2) (kha)

1869.5

1840.6 2.5 28.8

Slightly disturbed land (C ⫽ 0) (kha)

19.1

2.4 2.1 16.7

Land disturbance (kha)

Table 3. Land disturbance and greenhouse gas emissions caused by SWC

787.8

103.7 479.1 684.1

Greenhouse gas emissions (kt)

Assessing the Ecological Footprint of a Large Metropolitan Water Supplier 125

Repairs of motor vehicles, construction and other machinery Property developer, real estate, plant and vehicle hire and other property services Printing, stationery and services to printing Gas oil, fuel oil Insurance Sanitary and disposal services

Technical and computer services Automotive petrol Basic chemicals Road freight transport services Installation, repairs and maintenance of business equipment and computer hardware Communications services, revenue collection

Electricity supply Non-residential building and other construction Water supply, sewerage and drainage services

Sector

44 19 4 23 2

36 15 3 14 11

50

36

41

52

48

46

264 43 130 56

187

2 090

142 29 23 72

854

0.1

Degraded land (C ⫽ 0.8) (ha)

650

1

Consumed land (C ⫽ 1) (ha)

220 16 136 10

249

199

277

254

1 349 184 950 290

19

4 094

0.4

Replaced land (C ⫽ 0.6) (ha)

197 26 211 19

392

338

433

434

2 230 309 1 649 473

221

6 906

1

Significantly disturbed land (C ⫽ 0.4) (ha)

98 15 128 12

240

205

260

266

1 362 183 1 007 285

97

4 137

0.3

Partially disturbed land (C ⫽ 0.2) (ha)

404 46 202 13

404

344

446

410

2 167 541 1 519 504

1 840 633

7 121

1 234

Slightly disturbed land (C ⫽ 0) (ha)

261 28 224 29

426

375

467

468

2 327 334 1 558 537

2 359

7

1 191

Land disturbance (kt)

6 6 3 4

6

7

9

11

30 56 32 14

104

173

346

Greenhouse gas emissions (ha)

Table 4. Land disturbance and greenhouse gas emissions caused by SWC, by supplying sector

640 412 409 282

833

849

1 091

1 212

4 385 4 143 3 780 1 512

9 462

19 247

24 868

EF: land disturbance (ha)

126 M. Lenzen, S. Lundie, G. Bransgrove, L. Charet & F. Sack

Legal, accounting, marketing and business management services Typing, copying, mailing, cleaning, staff placement and other business services Newspapers, books, periodicals, recorded media and other publishing Miscellaneous manufacturing Footwear Health services, rehabilitation Air and space transport Natural gas Paints Electrical equipment Furniture Electronic equipment Construction machinery, material handling equipment Travel agencies, forwarding, storage, parking and other services to transport Gas production and distribution Education and training Taxi and hired car with driver

Fabricated metal products Clothing Accommodation, cafe´s and restaurants Banking Government administration, rates, registration, fines

10

5 10 13 6 1 0.4 2 2 5 2 1

1 0.1 1 0.2

7

5 4 0.4 3 5 0.2 1 2 1 2 1

1 0.4 0.4 1

9

5

9

16 11

2 6

8

11 42

9 1

5 1 3 1

4

44 41 56 27 8 1 16 9 20 8

54

57

56

132 68

50 131

8 1 5 1

7

45 75 112 44 11 2 24 15 38 13

86

84

79

178 101

88 283

5 1 3 1

4

25 46 71 27 7 1 14 9 22 8

53

51

47

112 61

54 179

8 1 5 2

9

71 64 67 37 12 6 27 19 42 14

83

90

87

115 96

88 168

9 2 6 2

8

58 76 104 47 17 2 25 16 36 15

92

93

87

187 109

94 262

0.3 0.2 0.0 0.1

0.3

1 1 0.1 1 1 1 0.5 1 0.2 0.4

1

1

2

1 1

3 0

27 16 13 6

31

139 124 113 85 64 60 57 55 49 44

183

191

196

230 205

280 275

Assessing the Ecological Footprint of a Large Metropolitan Water Supplier 127

128 M. Lenzen, S. Lundie, G. Bransgrove, L. Charet & F. Sack represents direct, on-site impacts, such as occupied land, catchments, fuel use and methane emissions. The next four commodities demonstrate the different relative importance of land disturbance and greenhouse gas emissions: the ratio of land to emission impacts (columns 7 and 8) is significantly higher for ‘technical and computer services’ and ‘basic chemicals’ than for ‘automotive petrol’ and ‘road freight transport services’, because the latter two commodities are associated with fuel combustion. Finally, ‘communication services’ and following lower-ranking positions represent less than 10% of the total EF.

Decomposition into Upstream Production Layers Applying the series expansion in equation (8) enables the decomposition of the land types and greenhouse gas emissions as in Tables 3 and 4 into production layers of increasing order (see Figure 2). Figure 2(a) shows that on-site consumed land represents about 50% of total requirements for consumed (mostly built) land. Including requirements of first-order suppliers, this figure increases to 75%. In other words, a typical audit-only analysis of SWC’s consumed land would hence result in about 75% of total land requirements, or 75% system completeness. While this figure is fairly acceptable, the situation is considerably different for other land types. Figure 2(b)–(e) show that for degraded, replaced, significantly and partially disturbed land requirements, 50% system completeness is only achieved after assessing suppliers up to third and fourth order. This is because these land types are typically used by agricultural industries and forestry, which do not appear as first- or second-order suppliers of SWC, but as higher-order suppliers in upstream input paths. However, the land impact at these higher orders can be significant, because agriculture and forestry are users of large tracts of land. Although SWC and its main first-order suppliers occupy mainly consumed land (C ⫽ 1), SWC’s main land impact occurs actually on significantly disturbed land (C ⫽ 0.4). This will become more evident when structural paths are appraised (section, ‘Structural path analysis’). Combining Figure 2(a)–(e), Figure 2(f) shows that for land disturbance, 50% system completeness is only achieved at about third order. The calculations carried out in this work employed a 135-sector input–output model, 42 of which appear as SWC’s first-order suppliers. Hence, in second order, there are 42 ⫻ 135 ⫽ 5670 possible input paths (compare the section above, ‘Structural path analysis’, after equation (9)). In third order, there are already 42 ⫻ 135 ⫻ 135 ⫽ 765 450 paths. This degree of complexity is generally beyond the scope of conventional EF assessments, due to limitations in time, research funds and human resources. As a result, project appraisals that do not employ input–output analysis will result in considerable systematic errors due to the truncation of the project’s system boundary. Finally, Figure 3 shows the convergence of SWC’s greenhouse gas emissions towards system completeness. On-site emissions (order ⫽ 0) are those originating from methane emissions and fuel combustion on SWC’s premises (about 120 kt carbon dioxide-e). First-order emissions are mainly those caused by fossilfuelled power plants supplying electricity to SWC (about 400 kt CO2-e). The remainder (about 280 kt CO2-e) is caused in upstream suppliers such as iron and steel producers. This will be illustrated in the following section.


129

Figure 2. SWC’s land requirements as a function of order of production layer: (a) C ⫽ 1; (b) C ⫽ 0.8; (c) C ⫽ 0.6; (d) C ⫽ 0.4; (e) C ⫽ 0.2; (f) land disturbance. Order 0 represents on-site effects, order 1 suppliers of SWC, order 2 suppliers of suppliers and so on.


Figure 3. Greenhouse gas emissions caused by SWC, as a function of production layer order. Order 0 represents on-site effects, order 1 suppliers of SWC, order 2 suppliers of suppliers and so on.

Structural Path Analysis A structural path extraction algorithm was run in order to obtain a decomposition of the land and emission requirements into structural paths (see equation (9)). In the following, each path will be characterized by a code, consisting of (1) a description of the path vertices (ij), (2) the path order, (3) the path value and (4) the path coverage, that is, the relative contribution (in %) to the total land type, emissions or EF. For the sake of brevity, vertex indices ij are assigned the codes listed in Table 5. For example, the path E Is Sm Nb (3; 0.4%), ranking 20th, denotes the land projected to become disturbed as a consequence of climate change, due to greenhouse gas emissions (E) caused by iron and steel plants (Is) supplying structural metal products (Sm) used by the non-residential building industry (Nb) to carry out contract work for SWC. The path is of third order, and constitutes a coverage of 0.4% of the total greenhouse gas emissions caused by SWC. The path value is 206 ha, which constitutes 0.3% of SWC’s EF. The reader should bear in mind that the values of these paths are only indicative, and that the ranking’s primary function is the identification and prioritization of targets for action on environmental impact abatement (Lenzen, 2001c). Hence, path values should not be interpreted as giving accurate figures for the absolute impact of SWC along particular supply chains. Table 6 shows that emissions (E) from power plants (El) are the most important component in SWC’s EF, contributing about 23 000 ha or 30.9%. On-site methane emissions (E Wa) contribute about 5400 ha or 7.3%, followed by on-site emissions from automotive petrol combustion (E Ap, 2.9%) and consumed land (C Wa, 2.9%). Subsequent paths are mostly emissions paths of first, second and third order, involving basic chemicals (Ch), non-residential building (Nb), road freight transport (Rd), electricity generation (El), iron and steel (Is), concrete (Cc), black coal mining (Bl) and disposal services (Gd), with intermittent land disturbance paths, representing consumed land for hydropower reservoirs and overhead lines across Australia (C El), degraded on-site land (D Wa), land consumed by non-residential building companies (C Nb) and land grazed


131

Table 5. Codes used for IOPC categories and land types in the vertex descriptions in Table 6 Symbol

IOPC; land types

Ap Bl C Cc Ce Ch Cp D E Ee El En Fo Gd Is Lm Nb Nf P Pd Rd Rh S Sg Sm Tp Ts Tx Wa Wo Wt

Automotive petrol Black coal Consumed land (C ⫽ 1) Concrete and mortar Cement Basic chemicals Plaster and other concrete products Degraded land (C ⫽ 0.8) Greenhouse gas emissions Cable, wire, batteries, lights and other electrical equipment Electricity Electronic equipment, photocopying, gaming machines Gas oil, fuel oil Sanitary and garbage disposal services Basic iron and steel, pipes, tubes, sheets, rods, bars and rails Lime Non-residential buildings, roads and other construction Non-ferrous metal recovery and basic problems Partially disturbed land (C ⫽ 0.4) Property developer, real estate and other property services Road freight transport services Repairs of household and business equipment Significantly disturbed land (C ⫽ 0.6) Sand, gravel and other construction materials mining Frames, mesh and other structural metal products Carpets, curtains, tarpaulins, sails, tents and other textiles Scientific research, technical and computer services Processed wool, textile fibres, yarns and woven fabrics Water supply, sewerage and drainage services Sheep and shorn wool Wholesale trade

by sheep supplying fibres for textiles products used in construction (Wo Tx Tp Nb). Remaining paths contribute less than 0.1% each to the total disturbancebased EF. Discussion Comparison with Previous Studies Input–output-based assessments of water supply and treatment have been carried out previously, with respect to energy requirements (Lough, 1996), their structural paths in a social accounting matrix (Xie, 2000) and in a life-cycle context (Hall et al., 2001). Lough (1996) finds energy requirements of about 1100 GJ per full-time employee for a US water provider, which is about equal to SWC’s value. There is so far only one EF case study that warrants comparison. The UK water provider Anglian Water Services has also used the EF as a tool to communicate environmental performance to stakeholders (Anglian Water Group, 2001), with first results based on an approach by Chambers & Lewis (2001). These authors calculate component EFs in a bottom-up manner, but with

132 M. Lenzen, S. Lundie, G. Bransgrove, L. Charet & F. Sack Table 6. Structural path analysis of SWC’s EF: E, land disturbance equivalent of greenhouse gas emissions; C, consumed; D, degraded; R, replaced; S, significantly disturbed Rank

Path

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39

E El (1; 42.1%) E Wa (0; 9.9%) E Ap (0; 3.9%) C Wa (0; 54.3%) E Ch (1; 2.9%) E Nb (1; 2.7%) E Is Nb (2; 1.1%) E Bl El (2; 1.0%) C El (1; 12.6%) E Rd (1; 0.7%) E El Nb (2; 0.7%) E Cc Nb (2; 0.7%) E Gd (1; 0.5%) E El Rh (2; 0.5%) C Nb (1; 5.3%) E El Ts (2; 0.4%) E Fo (0; 0.4%) E Ce Cc Nb (3; 0.4%) E Is Sm Nb (3; 0.4%) E El Ch (2; 0.3%) E El En Nb (3; 0.3%) D Wa (0; 9.4%) E Gd Nb (2; 0.2%) S Wo Tx Tp Nb (4; 4.2%) E Gd Ts (2; 0.2%) E Ce Cp Nb (3; 0.2%) S Wol Tsl (2; 1.6%) E El Cc Nb (3; 0.2%) E Lm Cc Nb (3; 0.2%) S Wa (0; 1.4%) R Wo Tx Tp Nb (4; 1.5%) E El Pd (2; 0.1%) E Wt Nb (2; 0.1%) E Rd Nb (2; 0.1%) E Fo (1; 0.1%) R Wo Ts (2; 1.2%) E Sg Nb (2; 0.1%) E El Is Nb (3; 0.1%) E Nf Ed Nb (3; 0.1%)

Area (ha)

Percentage of total

22 948 5 384 2 158 2 090 1 569 1 493 603 555 540 397 384 370 253 253 230 226 212 206 206 185 171 150 123 120 116 110 100 96 89 88 84 82 82 75 69 66 66 65 64

30.9 7.2 2.9 2.9 2.1 2.0 0.8 0.7 0.7 0.5 0.5 0.5 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.2 0.2 0.2 0.2 0.2 0.2 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1

the aim of capturing indirect effects in a life-cycle context (see Simmons & Chambers, 1998; Simmons et al., 2000; Barrett, 2001; Best Foot Forward, 2001). They deplore, however, a general lack of data, and deal with problems that are typical of a bottom-up auditing-type approach, such as boundary selection and double-counting. In addition, they consider only embodied energy, but not embodied land. Furthermore, the data underlying their case studies exhibit considerable discrepancies (Simmons et al., 2000; Chambers & Lewis, 2001), which could be due to the source data referring to different system boundaries (compare Lenzen & Munksgaard, 2001), and which make the results highly


133

data-dependent. Finally, and most importantly, these authors deal with a different quantity: appropriated bioproductivity rather than land disturbance. Their 1998–99 assessment of Anglian Water Services yielded an EF of 94 450 ha. This result is in the order of magnitude of SWC’s bioproductivity-based, first- and second-order EF (95 400 ha; see Table 2), but would be considerably higher if total requirements were taken into account (see Table 7). Furthermore, Anglian Water Services cause a smaller EF on a per-person basis, which is largely due to a much smaller per-person water consumption in the UK. In contrast, SWC exhibits a smaller EF per quantity of water supplied. It is interesting to see that for both providers, the largest shares of the total EF were due to grid electricity usage (Anglian Water Services 53% (Chambers & Lewis, 2001); SWC 30.9%, Table 6) for pumping stations, etc. Transportation (Anglian Water Services 4% (Chambers & Lewis, 2001); SWC 8%, Table 4) and waste (Anglian Water Services 3% (Chambers & Lewis, 2001); SWC 0.3%, Tables 4 and 6) constitute only minor parts. Finally, Chambers & Lewis (2001) do not list contributions from services provided for Anglian Water Services. This truncation causes a systematic error (see contributions from services to SWC in Table 4). Methodological and Analytical Shortcomings In addition to the uncertainty sources inherent in input–output analysis (see section above, ‘Errors in input–output-based EFs’), the analysis presented in the sections above, ‘Methodology’ and ‘Results’, suffers from the following shortcomings. First, the condition of parts of catchments managed by SWC or the Sydney Catchment Authority is unknown, and could therefore be a potential source of variations within SWC’s total EF. Note, however, that a substantial part of the land disturbance within the drinking water catchments has occurred as a consequence of land uses that are independent of SWC’s activities and operations. Some of these land uses, such as agriculture and coal mines, may actually be counted as part of SWC’s EF from higher-order production layers. Therefore, to include them as part of SWC’s footprint may lead to a case of double counting. It would perhaps be better then to only assess the land condition of catchment areas in direct ownership by the Sydney Catchment Authority and SWC. The condition of these catchments could be assessed by evaluating aerial photographs, complemented by site visits along with a review of the most recent catchment audit published by the Commonwealth Scientific and Industrial Research Organisation (CSIRO) for the Sydney Catchment Authority in December 2001. Secondly, critical parameters such as land conditions and conversion factors could be subjected to a sensitivity analysis, and the uncertainty calculus (see section above, ‘Errors in input–ouput-based EFs’) could be improved and extended to structural paths. Thirdly, some structural paths, particularly those of higher order, may describe average Australian supply chains, but not SWC’s specific supply chains, since some of SWC’s inputs may originate from particular suppliers that are atypical within the respective IOPC category in terms of their land use and emissions. Hence, some important structural paths could be further investigated by auditing the corresponding suppliers in a more detailed ‘manual’ process-chain-type analysis, thus refining the input–output model. Fourthly, procedures could be developed that allow EF credits obtained from beneficial reuses of effluents and land remediation.

351 495 624 880

9.4 4.1

Population served (106)

Note: Bioproductivity ⫽ appropriated bioproductivity.

Anglian Water Services SWC

Water supplied (ml/year) 37.4 156.2

Water consumption (kl/person/year) 0.27 0.15

Bioproductivity first and second order (ha/Ml) 0.21

Bioproductivity total (ha/Ml) 0.12

Land disturbance (ha/Ml)

EF

0.010 0.024

Bioproductivity first and second order (ha/person)

0.033

Bioproductivity total (ha/person)

0.018

Land disturbance (ha/person)

Table 7. Comparison of Anglian Water Services’ and SWC’s EFs (compare Chambers & Lewis (2001) and Table 2)



135

Finally, a major stumbling block in the usefulness of the EF to SWC as a strategic planning tool is its inability to incorporate ‘downstream’ impacts, which include the organization’s direct impacts on oceans and rivers. Management and Planning Relevance The calculation of SWC’s EF has enabled the organization to better understand and communicate its environmental performance in the provision of water, wastewater and stormwater services. It has also permitted greater transparency into some of its less observable impacts. Through subsequent calculations, the EF will allow SWC to gain a greater understanding in how it is progressing towards environmental sustainability through the determination of trends and changes across a number of different aspects of its environmental performance. Communication and Education The ability of the EF to aggregate a number of different aspects of the organization’s environmental performance into a single, easily understood indicator, and its usefulness as a communications tool, was seen by SWC as the main advantages in using the EF for corporate reporting, in particular its ability to engender a sense of personal responsibility amongst its customers in their use of water. For instance, by communicating the results of its pilot EF on a per-household and per-customer-served basis, in addition to the whole-corporation results, it is aimed to educate SWC’s customers, who play a significant role in the size of SWC’s footprint and hence its impacts on the environment. As an example, the use of scenario calculations proved to be a useful tool, where the results from EF calculations were used to illustrate that small changes in water consumption habits, such as the installation of water-saving showerheads and dual-flush toilets, can result in quite significant reductions in customers’ personal EF. These results support the key messages delivered by SWC’s demand management and water conservation programmes, which are aimed at achieving the ambitious target of a 35% reduction in per-person water usage by 2010 based on 1990 levels. Decision Making As Costanza (2000, p. 342) points out, the dangers of using a single-point indicator for decision making lie in the fact that “one can be ignorant of where the numbers came from, how they were aggregated”. It is for this reason that the decomposition and path extraction techniques described in the section above, ‘Methodology’, become valuable. The structural path analysis showed that by far the largest component of these off-site impacts were the greenhouse gas emissions emanating from the coalfired power stations supplying electricity to SWC. These emissions were responsible for nearly a third of the organization’s overall impact. This indicates that energy use should be a major consideration for SWC in its future planning and decision making. As a consequence, for SWC to make significant inroads into reducing the size of its EF, it will need to significantly reduce its requirement for energy in its operations or alternatively shift from its current reliance on using electricity sourced from the burning of fossil fuels. SWC has already taken steps to minimize the impacts associated with its use

136 M. Lenzen, S. Lundie, G. Bransgrove, L. Charet & F. Sack of energy and the generation of greenhouse gases from its sewage treatment plants. It is doing this through several different initiatives. It currently purchases 2.5% of its energy needs as ‘green power’ from the electricity grid and also generates around a further 4.2% of its overall energy needs as renewable energy through co-generation plants located at two of its sewage treatment plants. Investigations are also being undertaken to determine the feasibility of installing more co-generation plants at its sewage treatment plants as well as small-scale hydropower stations within the water distribution system. A detailed concept design report has also been completed for a proposed 5 MW hydro-electric facility to be located at the end of the pipeline that transfers water from Warragamba dam (supplying ⬃ 80% of Sydney’s water needs) to the Prospect water filtration plant. The conversion of sewage ‘biosolids’ into energy, utilizing waste to energy technologies, is a further initiative that SWC is investigating to minimize its current reliance on energy sourced from the burning of fossil fuels. It is also continually looking to improve the energy efficiency of its operations through the regular undertaking of audits at existing buildings, sewage treatment plants, water filtration plants, water pumping stations and sewage pumping stations that are responsible for a large proportion of the organization’s energy needs. The installation of energy-efficient motors at one of SWC’s water pumping stations has already resulted in a 10% reduction in energy usage. Energy monitoring and reporting procedures have also been established within SWC to allow the easier and more systematic identification of improvement opportunities in the organization’s use of energy. To accelerate the development of the potential renewable energy generation opportunities mentioned above, SWC is proposing to enter into a long-term strategic alliance with an energy service provider which will also assist the organization in developing an effective energy management programme. SWC has a large portfolio of assets and infrastructure. The structural path analysis showed that this land makes the third largest contribution to the footprint of SWC. Although the presence of concrete and metal infrastructure in the form of depots and administrative buildings makes this an inevitable contribution to the EF of SWC, there are lessons to be learnt for future strategic planning. Minimizing the disturbance of sites and the rehabilitation and restoration of degraded sites affected by SWC’s activities are two important initiatives that the organization can continue to take in future to minimize this component of its EF. The large contributions that other off-site impacts make to SWC’s overall EF indicate that SWC needs also to look carefully for opportunities to reduce its overall impact by modifying the inputs that its operations require. This may involve a reduction in the material intensity of its operations, where feasible from a public health and economic perspective, and where it will not affect the ecological and recreational amenity of waterways receiving treated wastewater discharge. Any concerted shift to the dematerialization of its operations to reduce its EF may involve specific investigations regarding suppliers’ use of resources, emissions, energy, waste and closed-loop recycling of materials. Further Improvements The EF is only one of many inputs into the strategic planning and management process. This is because the EF is only able to incorporate the ‘upstream’


137

environmental impacts of SWC’s operations, caused by the harvesting, manufacture and supply of the materials, energy and services that the organization needs to operate, along with the significant land area occupied by its vast portfolio of assets and infrastructure. Until the EF is capable of including SWC’s ‘downstream’ impacts (such as impacts on rivers and oceans), it will always be an underestimate of the total environmental impacts resulting from the organization’s activities. The EF will therefore only be one ‘driver’ of the organization’s planning towards sustainability, as it has the potential to neglect other environmental and social objectives of the organization. The need to consider downstream impacts and social factors in SWC’s decision making may require measures that lead to an increase in its EF, in order to provide appropriate levels of ecosystem protection or to meet community standards. An increase in energy consumption to improve wastewater treatment or to meet community standards for safe and reliable drinking water is a good example of this. Conclusions and Outlook The pilot EF study has allowed SWC to gain some valuable insights into the impacts associated with its current operations and progress towards sustainability. In particular it has shown that the organization’s use of energy and the impacts associated with its supply chain make a significant contribution to SWC’s overall EF. The EF also appears very promising in terms of its use as an education and communication tool. It may also have potential as a planning and decision support tool. As it continues to gain international acceptance as an indicator of sustainability, the EF also shows potential as a valuable benchmarking tool, allowing comparisons of sustainability performance between and within industrial sectors. This will assist organizations like SWC to better communicate what performance against ESD indicators means in terms of cumulative impacts and overall progress towards ESD. The benchmarking potential of the EF is flagged by some interesting comparisons between Anglian Water Services and SWC, including an apparent ecological efficiency in SWC’s processes. Further research would be required, however, to verify this indicative result. Despite the obvious benefits of the EF as an indicator of sustainability, a number of methodological issues have become apparent that limit the use of the EF as a ‘stand-alone’ environmental and sustainability planning tool for SWC. The inability of the EF to consider the ‘downstream’ impacts of the organization’s activities, and the limited type of sustainability indicators capable of being included in the EF, means that the EF will not be a true reflection of sustainability performance. For this reason the EF can only be one input into the organization’s environmental planning and decision-making processes, supplementing the extensive environmental monitoring and community and stakeholder consultation that are currently other significant inputs. The ability of the EF to incorporate the downstream impacts of SWC’s operations through comprehensive research would have significant benefits to SWC, and also other water services providers, in terms of assessing and communicating the organization’s progress towards sustainability. The inclusion of these downstream impacts would enable a better comparative analysis of the less observable off-site impacts caused by the inputs into SWC’s operations. This

138 M. Lenzen, S. Lundie, G. Bransgrove, L. Charet & F. Sack would assist future decision making and planning within SWC by ensuring that the cumulative environmental impacts resulting from the organization’s activities could be considered against specific aspects of the organization’s environmental performance. This would undoubtedly result in improved environmental and social outcomes through a more holistic, integrated analysis of future options, and help the organization in achieving its goal of becoming a sustainable water services provider. Notes 1.

2. 3.

Barrett (2001) and Simmons et al. (2000) examine regional, component-based EFs using local data, and considering some upstream impacts such as from car manufacturing. The present authors, however, could not determine the boundary conditions of their life-cycle approach, and hence the degree of completeness of the figures. CO2-e: carbon dioxide equivalents. The authors use the fuelwood equivalence approach for converting energy into land.

References Anglian Water Group (2001) The Ecological Footprint and Anglian Water Group (http://www. bestfootforward.com/FootprintNews/Newsletter files/Anglianattach.htm) (Best Foot Forward, Future Centre, 115 Magdalen Road, Oxford OX4 1RQ, UK). Arrow, K., Bolin, B., Costanza, R., Dasgupta, P., Folke, C., Holling, C.S., Jansson, B.-O., Levin, S., Ma¨ler, K.G. & Perrings, C. (1995) Economic growth, carrying capacity, and the environment, Science, 268, pp. 520–521. Australian Bureau of Statistics (1999a) Australian National Accounts, Input–Output Tables, 1994–95, Australian Bureau of Statistics Catalogue No. 5209.0 (Canberra, Australian Bureau of Statistics). Australian Bureau of Statistics (1999b) Australian National Accounts, Input–Output Tables, Product Details, 1994–95, Australian Bureau of Statistics Catalogue No. 5215.0 (Canberra, Australian Bureau of Statistics). Ayres, R.U. (2000) Commentary on the utility of the ecological footprint concept, Ecological Economics, 32, pp. 347–349. Barrett, J. (2001) Component ecological footprint: developing sustainable scenarios, Impact Assessment and Project Appraisal, 19(2), pp. 107–118. Bergh, J.C.J.M. van den & Verbruggen, H. (1999) Spatial sustainability, trade and indicators: an evaluation of the ‘ecological footprint’, Ecological Economics, 29(1), pp. 61–72. Best Foot Forward (2001) Ecological Footprint Calculators (http://www.bestfootforward.com) (Best Foot Forward, Future Centre, 115 Magdalen Road, Oxford OX4 1RQ, UK). Best Foot Forward & Imperial College (2001) Ecological Footprint of the Isle of Wight (http://www. bestfootforward.com/island form.htm) (Best Foot Forward, Future Centre, 115 Magdalen Road, Oxford OX4 1RQ, UK). Bicknell, K.B., Ball, R.J., Cullen, R. & Bigsby, H.R. (1998) New methodology for the ecological footprint with an application to the New Zealand economy, Ecological Economics, 27(2), pp. 149– 160. Bullard, C.W. & Sebald, A.V. (1977) Effects of parametric uncertainty and technological change on input–output models, Review of Economics and Statistics, 59, pp. 75–81. Bullard, C.W. & Sebald, A.V. (1988) Monte Carlo sensitivity analysis of input–output models, Review of Economics and Statistics, 70(4), pp. 708–712. Bullard, C.W., Penner, P.S. & Pilati, D.A. (1978) Net energy analysis—handbook for combining process and input–output analysis, Resources and Energy, 1, pp. 267–313. Chambers, N. & Lewis, K. (2001) Ecological Footprint Analysis: Towards a Sustainability Indicator for Business, Research Report No. 65 (London, Association of Chartered Certified Accountants). Close, A. & Foran, B. (1998) Canberra’s Ecological Footprint, Working Paper Series 98/12 (Canberra, Resource Futures Program, CSIRO Wildlife and Ecology). Commonwealth of Australia (1992) National Strategy for Ecologically Sustainable Development (Canberra, Government Publishing Service). Costanza, R. (2000) The dynamics of the ecological footprint concept, Ecological Economics, 32, pp. 341–345.


139

Crama, Y., Defourny, J. & Gazon, J. (1984) Structural decomposition of multipliers in input–output or social accounting matrix analysis, Economie Applique´e, 37, pp. 215–222. Darwin, R., Tsigas, M., Lewandrowski, J. & Raneses, A. (1996) Land use and cover in ecological economics, Ecological Economics, 17, pp. 157–181. Defourny, J. & Thorbecke, E. (1984) Structural path analysis and multiplier decomposition within a social accounting matrix framework, Economic Journal, 94, pp. 111–136. Dixon, R. (1996) Inter-industry transactions and input–output analysis, Australian Economic Review, 3(115), pp. 327–336. Dobben, H.F. van, Schouwenberg, E.P.A.G., Nabuurs, G.J. & Prins, A.H. (1998) Biodiversity and productivity parameters as a basis for evaluating land use changes in LCA, in: IVAM Environmental Research (Ed) Biodiversity and Life Support Indicators for Land Use Impacts, Publication Series Raw Materials Nr 1998/07 (Delft, Ministry of Transport, Public Works and Water Management). Duchin, F. (1992) Industrial input–output analysis: implications for industrial ecology, Proceedings of the National Academy of Science of the USA, 89, pp. 851–855. Ferng, J.-J. (2002) Toward a scenario analysis framework for energy footprints, Ecological Economics, 40, pp. 53–69. Flint, K. (2001) Institutional ecological footprint analysis, International Journal of Sustainability in Higher Education, 2(1), pp. 48–62. Glynn, J. & Sack, F. (2001) Stakeholder and Sydney Water requirements for ecologically sustainable development (ESD) indicators, Platform Paper Sustainability 2001—The Way Forward Proceedings, available at: http://members.iinet.net.au/ ⬃ conswa/. Graetz, R.D., Wilson, M.A. & Campbell, S.K. (1995) Landcover Disturbance over the Australian Continent, Biodiversity Series, Paper No. 7 (Canberra, Department of the Environment, Sport and Territories Biodiversity Unit). Hall, M., Lovell, H., White, S., Lenzen, M. & Mitchell, C. (2001) Management of Risk for On-site and Centralised Wastewater Treatment, Case Study Berowra Creek (Sydney, Hornsby Shire Council). Hobbs, R.J. & Hopkins, A.J.M. (1990) From frontier to fragments: European impact on Australia’s vegetation, Proceedings of the Ecological Society of Australia, 16, pp. 93–114. Hondo, H. & Sakai, S. (2000) Preliminary life cycle inventory analysis (Pre-LCI) using an economic input–output table, in: The Fourth International Conference on EcoBalance (Tsukuba, National Research Institute for Metals). Intergovernmental Panel on Climate Change (1995) Climate Change 1995: Scientific–Technical Analysis of Impacts, Adaptations, and Mitigation of Climate Change. Summary for Policymakers, Technical Summary of the Working Group II Report (Bracknell, Intergovernmental Panel on Climate Change). Isard, W., Bassett, K., Choguill, C., Furtado, J., Izumita, R., Kissin, J., Romanoff, E., Seyfarth, R. & Tatlock, R. (1967) On the linkage of socio-economic and ecologic systems, Papers and Proceedings of the Regional Science Association, 21, pp. 79–99. Joshi, S. (2001) Product environmental life-cycle assessment using input–output techniques, Journal of Industrial Ecology, 3(2–3), pp. 95–120. Ko¨llner, T. (2000) Species-pool effect potentials (SPEP) as a yardstick to evaluate land-use impacts on biodiversity, Journal of Cleaner Production, 8(4), pp. 293–311. Kooten, G.C. van & Bulte, E.H. (2000) The ecological footprint: useful science or politics?, Ecological Economics, 32, pp. 385–389. Lave, L.B., Cobas-Flores, E., Hendrickson, C.T. & McMichael, F.C. (1995) Using input–output analysis to estimate economy-wide discharges, Environmental Science & Technology, 29(9), pp. 420A–426A. Lenzen, M. (2001a) Errors in conventional and input–output-based life-cycle inventories, Journal of Industrial Ecology, 4(4), pp. 127–148. Lenzen, M. (2001b) A generalised input–output multiplier calculus for Australia, Economic Systems Research, 13(1), pp. 65–92. Lenzen, M. (2002) A guide for compiling inventories in hybrid LCA: some Australian results, Journal of Cleaner Production, 10, pp. 545–572. Lenzen, M. & Munksgaard, J. (2001) Energy and CO2 analyses of wind turbines—review and applications, Renewable Energy, 26(3), pp. 339–362. Lenzen, M. & Murray, S.A. (2001) A modified ecological footprint method and its application to Australia, Ecological Economics, 37(2), pp. 229–255. Leontief, W. & Ford, D. (1970) Environmental repercussions and the economic structure: an input–output approach, Review of Economics and Statistics, 52(3), pp. 262–271. Levett, R. (1998) Footprinting: a great step forward, but tread carefully, Local Environment, 3(1), pp. 67–74.

140 M. Lenzen, S. Lundie, G. Bransgrove, L. Charet & F. Sack Lewis, K., Simmons, C. & Chambers, N. (2000) An Ecological Footprint Analysis of Different Packaging Systems (Oxford, Best Foot Forward). Lindeijer, E. (2000) Review of land use impact methodologies, Journal of Cleaner Production, 8, pp. 273–281. Lough, T.S. (1996) Energy analysis of the structures of industrial organizations, Energy, 21(2), pp. 131–139. Moffatt, I. (2000) Ecological footprints and sustainable development, Ecological Economics, 32, pp. 359– 362. National Greenhouse Gas Inventory Committee (1998) Australian Methodology for the Estimation of Greenhouse Gas Emissions and Sinks, Workbooks 1.1 to 8.1 (Canberra, Australian Greenhouse Office). Opschoor, H. (2000) The ecological footprint: measuring rod or metaphor?, Ecological Economics, 32, pp. 363–365. Quandt, R.E. (1959) On the solution of probabilistic Leontief systems, Naval Research Logistics Quarterly, 6(4), pp. 295–305. Rapport, D.J. (2000) Ecological footprints and ecosystem health: complementary approaches to a sustainable future, Ecological Economics, 32, pp. 381–383. Rees, W.E. (1992) Ecological footprints and appropriated carrying capacity: what urban economics leaves out, Environment and Urbanization, 4(2), pp. 121–130. Rees, W. (1996) Revisiting carrying capacity: area-based indicators of sustainability, Population and Environment, 17(3), pp. 195–215. Rees, W.E. (2001) Ecological footprint, concept of, in: S. A. Levin (Ed) Encyclopedia of Biodiversity (New York, Academic Press). Rees, W. & Wackernagel, M. (1996) Urban ecological footprints: why cities cannot be sustainable— and why they are a key to sustainability, Environmental Impact Assessment Review, 16(4–6), pp. 223–248. Simmons, C. & Chambers, N. (1998) Footprinting UK households: how big is your ecological garden?, Local Environment, 3(3), pp. 355–362. Simmons, C., Lewis, K. & Barrett, J. (2000) Two feet—two approaches: a component-based model of ecological footprinting, Ecological Economics, 32, pp. 375–380. Simpson, R., Petroeschevsky, A. & Lowe, I. (1998) The Ecological Footprint of Australia, with a Focus on the South-east Queensland Region, Preliminary Technical Report (Logan Campus, School of Public Health, Griffith University). Simpson, R.W., Petroeschevsky, A. & Lowe, I. (2000) An ecological footprint analysis for Australia, Australian Journal of Environmental Management, 7, pp. 11–18. Suh, S. & Huppes, G. (forthcoming) Techniques for life cycle inventory of a product, Journal of Cleaner Production, in review. Swan, G. & Pettersson, B. (1998) Land use evaluation in forestry, in: G. Swan (Ed) Evaluation of Land Use in Life Cycle Assessment, Center for Environmental Assessment of Product and Material Systems, CPM Report 1998⬊2 (Go¨teborg, Chalmers University of Technology). Sydney Water (2001a) Public Consultation Report on Indicators for Ecologically Sustainable Development, available at: http://www.sydneywater.com.au/html/Environment/pdf/ESD Report (final).pdf. Sydney Water (2001b) Towards Sustainability Report 2001, available at: http:/www.sydneywater. com.au/html/tsr/index.html. Treloar, G. (1997) Extracting embodied energy paths from input–output tables: towards an input– output-based hybrid energy analysis method, Economic Systems Research, 9(4), pp. 375–391. Treloar, G.J. (1998) A comprehensive embodied energy analysis framework, Ph.D. dissertation (Geelong, Deakin University). Treloar, G., Love, P.E.D., Iyer-Raninga, U. & Faniran, O.O. (2000) A hybrid life cycle assessment method for construction, Construction, Management and Economics, 18, pp. 5–9. Venetoulis, J. (2001) Assessing the ecological impact of a university, International Journal of Sustainability in Higher Education, 2(2), pp. 180–196. Vuuren, D.P. van & Smeets, E.M.W. (2000) Ecological footprints of Benin, Bhutan, Costa Rica, and the Netherlands, Ecological Economics, 34, pp. 115–130. Wackernagel, M. (1998) The ecological footprint of Santiago de Chile, Local Environment, 3(1), pp. 7–25. Wackernagel, M. & Rees, W. (1995) Our Ecological Footprint: Reducing Human Impact on the Earth (Philadelphia, PA, New Society Publishers). Wackernagel, M., Onisto, L., Linares, A.C., Falfan, I.S.L., Garcia, J.M., Guerrero, A.I.S. & Guerrero, M.G.S. (1997) The Ecological Footprint of Nations, Study for the Rio ⫹ 5 Forum (Xalapa, Centro de Estudios para la Sustentabilidad).


141

Wackernagel, M., Onisto, L., Bello, P., Linares, A.C., Falfan, I.S.L., Garcia, J.M., Guerrero, A.I.S. & Guerrero, M.G.S. (1999) National natural capital accounting with the ecological footprint concept, Ecological Economics, 29, pp. 375–390. Wackernagel, M., Monfreda, C. & Gurarie, E. (2002a) Improvements to National Footprint Accounts since the Living Planet Report 2000, available at: www.redefiningprogress.org (Oakland, CA, Redefining Progress). Wackernagel, M., Schulz, N.B., Deumling, D., Callejas Linares, A., Jenkins, Kapos, V., Monfreda, C., Loh, J., Myers, N., Norgaard, R. & Randers, J. (2002b) Tracking the ecological overshoot of the human economy, Proceedings of the National Academy of Sciences, 99(14), pp. 9266–9271. Wood, R. & Lenzen, M. (in press) An application of an improved ecological footprint method and structural path analysis in a comparative institutional study, Local Environment (in press). Xie, J. (2000) An environmentally extended social accounting matrix, Environmental and Resource Economics, 16, pp. 391–406.