EC O L O G IC A L E C O N O M IC S 5 9 ( 2 0 06 ) 11 5 –1 30

                                               a v a i l ...
116                                         EC O LO GIC A L E CO N O M ICS 5 9 ( 2 00 6 ) 1 1 5 –1 30

EC O L O G IC A L E C O N O M IC S 5 9 ( 2 0 06 ) 11 5 –1 30                               117

118                                          EC O LO GIC A L E CO N O M ICS 5 9 ( 2 00 6 ) 1 1 5 –1 30

    Many aspect...
EC O L O G IC A L E C O N O M IC S 5 9 ( 2 0 06 ) 11 5 –1 30                                         119

 Table 1 – Curr...
120                                     EC O LO GIC A L E CO N O M ICS 5 9 ( 2 00 6 ) 1 1 5 –1 30

EC O L O G IC A L E C O N O M IC S 5 9 ( 2 0 06 ) 11 5 –1 30                                 121

most relevant subject t...
122                                       EC O LO GIC A L E CO N O M ICS 5 9 ( 2 00 6 ) 1 1 5 –1 30

cator of populatio...
EC O L O G IC A L E C O N O M IC S 5 9 ( 2 0 06 ) 11 5 –1 30                                       123

analysis, which c...
124                                            EC O LO GIC A L E CO N O M ICS 5 9 ( 2 00 6 ) 1 1 5 –1 30

EC O L O G IC A L E C O N O M IC S 5 9 ( 2 0 06 ) 11 5 –1 30                                125

126                                                                                   EC O LO GIC A L E CO N O M ICS 5 9 (...
EC O L O G IC A L E C O N O M IC S 5 9 ( 2 0 06 ) 11 5 –1 30                                  127

   In the life cycle a...
Ecological –economic modelling for strategic regional waste ...
Ecological –economic modelling for strategic regional waste ...
Ecological –economic modelling for strategic regional waste ...
Upcoming SlideShare
Loading in …5

Ecological –economic modelling for strategic regional waste ...


Published on

Published in: Business, Technology
1 Like
  • Be the first to comment

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Ecological –economic modelling for strategic regional waste ...

  1. 1. EC O L O G IC A L E C O N O M IC S 5 9 ( 2 0 06 ) 11 5 –1 30 a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m w w w. e l s e v i e r. c o m / l o c a t e / e c o l e c o n ANALYSIS Ecological–economic modelling for strategic regional waste management systems S.E. Shmelev a,⁎, J.R. Powell b a EERU, The Open University, UK b CCRU, University of Gloucestershire, UK AR TIC LE I N FO ABS TR ACT Article history: This paper summarises some recent work exploring the development of a multi-criteria Received 17 September 2004 optimisation tool for achieving sustainable solutions for municipal solid waste management Received in revised form systems (MSWMS). The aim of the project was to provide a new methodological background 23 September 2005 for the regional solid waste management modelling taking into account spatial and Accepted 27 September 2005 temporal patterns of waste generation and processing, environmental as well as economic Available online 5 December 2005 impacts of the system development with a particular emphasis on public health and biodiversity. Keywords: The research has focused on integrating three different approaches to the spatial-temporal Ecological–economic modelling analysis of the MSWMS, namely a life cycle inventory analysis, which helps to identify Waste management emission patterns within the MSWMS, a multi-criteria optimisation approach, which helps Integrated approach to find compromise solutions among environmentally and economically preferred options, UK and a geographic information systems approach, which provides a tool for identifying waste management facilities, transportation environmental and social impacts, as well as analysis of environmental impacts on valuable ecosystems. A Russian methodology for calculating environmental damage was used to weight the importance of different sub-territories covered by the system as well as simplifying the analysis of emissions from the waste treatment plants. The approach provides a new perspective for the analysis of municipal solid waste management systems at the regional scale. The principal novelty of the proposed complex MSW strategic management model is an integration of the different types of data–geographical, environmental and economic–using relational database technology. Simulations using the dataset for Gloucestershire were performed on a simplified version of the model. Simulations were undertaken to explore the potential effects on waste management infrastructure of introducing the EU Landfill Directive. Understanding the strengths and weaknesses inherent in the methods utilised has suggested that a relatively affordable and easy to use tool can be developed for strategic analysis of the municipal solid waste management system in a region, giving useful support to the decision-maker regarding the potential development paths and trade-offs between economic and environmental performance of a proposed waste management system. © 2005 Elsevier B.V. All rights reserved. ⁎ Corresponding author. Tel.: +44 7729733366. E-mail address: (S.E. Shmelev). 0921-8009/$ - see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.ecolecon.2005.09.030
  2. 2. 116 EC O LO GIC A L E CO N O M ICS 5 9 ( 2 00 6 ) 1 1 5 –1 30 ecological economic modelling approach that attempts to 1. Introduction integrate life cycle inventory analysis, environmental impact assessment and economic appraisal within a geo- Strategic decision-making for dealing with municipal solid graphic information system (GIS) framework. The aim is not waste is a problem currently exercising the minds of many to provide an “optimum” solution but to highlight to deci- local governments (Gloucestershire County Council, 2002) sion-makers the trade-offs inherent through investing in throughout the European Union (EU). This paper is devoted different mixes of waste management technology at a to ecological–economic modelling of the strategic develop- range of scales from the local to the regional. In other ments in Municipal Solid Waste (MSW) Management Systems words, it can reveal, for a particular area or region, how at the regional level. waste management should be ‘integrated’ in order to The waste management problem in the EU is characterised achieve the BPEO solution. by increasing per capita production of waste materials, the The waste management problem has a complex nature need for high levels of investment in physical infrastructure with a range of important dimensions such as multiplicity of (incinerators, landfills, recycling facilities), institutional bar- the types of waste generated in the system, complex spatial riers (such as the long-term nature of contracts), a wide range pattern of waste arisings, the necessity to transport waste of stakeholders and a dynamic policy arena (e.g. the Waste long distances for processing, a variety of emissions from Electrical and Electronic Equipment and Landfill Directives are waste collection, transporting and treatment to the environ- two instruments aimed at reducing the amounts of biode- ment, and the almost unpredictable and localised character of gradable and electronic waste being landfilled). The waste impacts of these emissions on humans and ecosystems. And stream itself varies in composition over time and space with although there have been attempts to analyse regional waste seasonal and longer term changes in the quantities and management systems taking into account environmental amounts of various materials and the market for ‘recycled’ impacts of processes under study, most of them have not materials is characterised by uncertain demand and fluctuat- formed a holistic method for analysing all spatial, temporal ing prices. as well as qualitative aspects of the problem. Therefore, the Strategic decision-making for waste is a complex pro- aim of the paper is to provide a new methodological back- blem that appears to offer scope to mathematical model- ground developing regional municipal solid waste manage- ling procedures in order to find “optimal” solutions. ment modelling, taking into account spatio-temporal pat- Although standard modelling approaches are limited as terns of waste generation and processing, environmental as the ideal solution looks very different depending where well as economic impacts of the system development with a you are situated: from the household or local government particular emphasis on public health and biodiversity. point of view the best solution would be to eliminate the This paper takes the first steps to develop a model for waste and remove the need for any waste service provision municipal solid waste management system at the regional in the first place, while the view from the waste industry level. The paper analyses the post-consumption stages of would be one of maximising the number of waste streams the waste life cycle, namely collection, sorting, treatment and quantities of waste over time to ensure survival of the and final disposal. The municipal solid waste management industry. This paper considers whether ecological economic system under study is illustrated by Fig. 1, which shows the modelling approach has anything new to offer the main material flows within the system. The figure reveals policymaker. that the whole life cycle of materials entering and leaving the waste management system consists of several stages— raw materials extraction, processing, sale, consumption, 2. Description of the problem finally becoming waste when they are discarded by consu- mers. These materials in the waste stream then undergo col- Ecological–economic modelling is an aid to strategic deci- lection, sorting (removal of recyclable materials) and sion-making for waste management where there is nearly treatment (which can be thermal or biological), with the final always strong local opposition to the siting of waste facil- stage being disposal in the landfill. The shaded areas in the ities, where alternative waste management approaches diagram are the stages of the life cycle explicitly taken place heavy demands on the environment, where future account of in this paper. EU policy threatens to put the onus on producer responsi- bility and thus remove significant quantities of high value materials from the waste stream, and solutions are driven 3. Previous and current approaches to waste as much by local politics as by economic factors. Standard management modelling economic modelling approaches seeking the optimum or least cost solution fail as they cannot incorporate the There have been many attempts to analyse municipal solid wide range of factors that need to be included in a decision waste management systems over the past decade. Economic that must be based on achieving the Best Practicable Envir- as well as environmental and social aspects of their perfor- onmental Option (BPEO). Decision-makers do need assis- mance have been taken into account. Despite the large tance in making strategic choices that cause social and amount of research done, the application of the major meth- environmental impacts, and tie up large amounts of ods employed does not provide a holistic picture of munici- money and land for significant periods of time. The pal solid waste management systems that can examine approach presented here is a first step in developing an environmental impacts and the economic costs of siting,
  3. 3. EC O L O G IC A L E C O N O M IC S 5 9 ( 2 0 06 ) 11 5 –1 30 117 System boundary Raw materials extraction Impacts of processes Impacts of transport Unseparated waste Non-organic waste Processing Fully separated waste Organic waste Recovered energy Recovered material Sale Stages of the waste management process Consumption Waste collection Deep Organic/ Without separation non-organic separation Paper, glass, Sorting metals, organic etc. Waste treatment Incineration Complex recycling Composting Landfill Fig. 1 – The municipal solid waste management system: material flows. technological processes involved, transportation, impacts (1997) looked at management costs, air pollution and the and their spatio-temporal distribution, or identify the parties recycling goals, but missed out water and soil pollution, affected. The main spheres of research in the field of noise, road congestion, employment and health impacts; MSWMS in 1990s have been: analysis of waste generation Haastrup et al. (1998) concentrated on costs, air, water and determinants (Hockett et al., 1995; Daskalopoulos et al., soil pollution, road congestion, technological reliability, but 1998; Chen and Chang, 2000, siting of waste management did not cover noise, employment, health impacts and recy- facilities (Huang et al., 1995; Chang and Wang, 1996; Fre- cling goals. driksson, 2000), the choice of the waste treatment method A substantial amount of research on local aspects of muni- (Huhtala, 1997; Dalemo et al., 1998; Highfill and McAsey, cipal solid waste management modelling has been carried out 2001), environmental impacts of different waste manage- using LCI methodology based on the recent models developed ment technologies (Nixon et al., 1997; Slater and Frederick- by White et al. (1999) and the Environment Agency's WISARD son, 2001; Powell, 1996), economic mechanism of managing model. Powell et al. (1996), for example, compared environ- MSWS (Morris and Holthausen, 1994; Jenkins et al., 2000; mental and social impacts of a kerbside collection scheme for Palmer et al. 1997; Fullerton and Wu, 1998; Hong, 1999), recyclable household waste with a bring scheme, using life transportation of waste (Bhat, 1996; Kulcar, 1996), macroeco- cycle assessments and economic valuation for assigning rela- nomics of recycling (Nakamura, 1999; Ferrer and Ayres, 2000; tive weights to these impacts, while Powell et al. (1998) Masui et al. 2000) and complex planning (Huang et al., 1997; explored alternative approaches to waste management for Chang and Wang, 1997; Chang et al., 1997; Haastrup et al., six district councils in Gloucestershire. Powell (2000) investi- 1998). In the majority of this research, the focus has been on gated the potential for using LCI analysis in local authority single aspects of the problem, for example, Chang and Wang waste management decision-making.
  4. 4. 118 EC O LO GIC A L E CO N O M ICS 5 9 ( 2 00 6 ) 1 1 5 –1 30 Many aspects of the waste management systems perfor- Hyun, 1999; Craighill and Powell, 1996; Powell, 2000), opera- mance were not integrated in a holistic model taking into tions research methods (Chang and Wang, 1997; Chang et account spatial distribution of environmental as well as al., 1997), multi-criteria assessment (Hokkanen and Salmi- economic impacts, and analysing transportation, technolo- nen, 1997; Rogers and Bruen, 1998; Salminen et al., 1998) and gical and siting issues simultaneously. Unfortunately, expert systems (Barlishen and Baetz, 1996; Haastrup et al., almost all of these models are of minimal use by the deci- 1998). All of these methods have particular uses in specific sion-makers as they miss some of the key institutional areas and Table 1 below identifies their strengths and dimensions of waste management as identified by Vigileos weaknesses. (2002), in particular the unequal social impacts of waste The life cycle inventory approach (see Fig. 2) provides management, the nature of contracts drawn up between information on the spectrum and quantities of emissions the industry and local authorities which are long-term, from a given technological process and when it comes to and sometimes require delivery of guaranteed amounts of comparing different scenarios sophisticated methods of waste, political pressures to recycle, barriers imposed by multi-criteria assessment (Munda and Romo, 2001) could be government regulations and the lack of communications applied. However, LCI methodology does not include any geo- between different participants in the waste management graphical or time dimension nor provide any estimates of the sector. effect of the emissions inventoried. When used in isolation, it Although there are examples of environmental–economic cannot identify the best solution (i.e. BPEO) of the waste man- analysis of municipal solid waste management systems on agement problem. the regional level by Haastrup et al. (1998), Chang and Lin MCDA, optimisation, Delphi on the other hand allow for (1997) and Chang and Wang (1996), many applications do comparison between alternatives that need to be integrated not incorporate an integrated analysis of environmental with an approach that can analyse the waste management impacts from all stages of the life cycle of municipal solid system itself. waste, spatial ecological–economic modelling of the distri- Geographic Information technology is a powerful tool for bution of impacts, non-substitutable treatment of environ- analysing and exhibiting spatial data. However, rating and mental and economic characteristics of the development of scoring of several scenarios (which is done often in geo- the system, including non-monetary valuation of environ- spatial environmental impact assessment, EIA) is not mental damage. What is missing is a technique for solving enough for performing an integrated analysis of the devel- regional waste problems which inevitably have a large num- opment of the municipal solid waste management system. ber of possible solutions due to variable population densi- It is necessary to perform a significant amount of simula- ties, incomes, multiple (actual and potential) locations for tion experiments, changing different spatial siting patterns, waste management infrastructure, protected landscape processing capacities, waste collection and sorting schemes areas and high value ecological sites. There is thus an to arrive at the decision space from which a selection can urgent need for improved methods for identifying BPEO1 be made. All of these approaches need to be underpinned solutions to waste management problems at the regional by some impact assessment methodology. The one selected level. The range of potential development paths for a solid here is the Russian methodology for environmental damage waste management system, for example, could include a calculation that was developed by Balatsky et al. (Vremen- large centralised regional facility, or a set of small localised naja tipovaja metodika, 1983; Vremennaja metodika, 1999) ones, depending on the physical conditions, and this re- and allows for taking into account the spatial dimension of presents a situation of choice between multidimensional environmental impacts in the form of coefficients of envir- scenarios. onmental value of the territories or regions (Vremennaja metodika, 1999). At the same time, it lowers the dimension of the analysed vector of environmental characteristics of 4. A comparison of approaches for analysing the given waste management system, which can be in turn the municipal solid waste management problem divided into negative effects of recycling, incineration and other waste treatment options, as well as negative effects Among the methods used for analysis of MSWMS during the on air, water and soil. Such lowering of the dimension past 10 years several should be mentioned here: input–out- simplifies the decision-making significantly and allows for put approach (Nakamura, 1999; Ferrer and Ayres, 2000), mul- dealing with only two dimensions of the waste manage- tiple regression analysis (Hockett et al., 1995; Daskalopoulos ment planning problem—environmental and economic. et al., 1998), life cycle analysis (Powell et al., 1998; Song and The Russian methodology for estimating environmental damage uses coefficients of environmental harm, attributed to each type of emission into water and air. These coeffi- 1 The standard notion of BPEO-Best Practicable Environmental cients are developed from laboratory based biological Option is defined as follows “A BPEO is the outcome of a research on animals (i.e. standard toxicological studies) systematic and consultative decision making procedure, which and extrapolation of these effects on humans. This informa- emphasises the protection and conservation of the environment tion is then integrated with another set of coefficients— across land, air and water. The BPEO procedure establishes, for a coefficients of environmental value of the territories or given set of objectives, the option that provides the most benefits of least damage to the environment as a whole, at acceptable regions that are based on the ecosystem value of the cost, in the long term as well as the short term”, in the 12th Report major biomes, soils, water reserves, located in the territory of the Royal Commission on Environmental Pollution, 1988. of the given region.
  5. 5. EC O L O G IC A L E C O N O M IC S 5 9 ( 2 0 06 ) 11 5 –1 30 119 Table 1 – Current tools for municipal solid waste management decision-making Method Strengths Weaknesses LCI—life cycle inventory • Reflects a wide spectrum of emissions • Only an inventory of emissions • Allows integration of environmental data with economic • No information on impacts to the recipients data • No time or space related dimensions • Flexible, allows easy comparison of different scenarios • Unable to make local/regional/global trade-offs MCDA—multi-criteria • Allow comparison of multi-attribute or multi-objective • Problems with weight estimation decision analysis scenarios • Limitations by comparing only a relative small • Flexibility in the choice of criteria number of alternatives, which could not represent • Allows integration of quantitative and qualitative data the efficient set of solutions Optimisation • Gives the best solution from the feasible set • The opportunity to solve large scale non-linear • Permits solving of multi-objective problems by mixed integer problems limited by the existing employing goal programming, compromise programming algorithms techniques, etc. • Certain assumptions about the relationships in • Allows the user to identify the efficient frontier of the the model have to be made solution space for subsequent decision-making GIS—geo-information • Reflects spatial patterns of the geographical distribution • Does not have a time dimension systems of actors, flows and sensitive areas • Requires integration with other techniques for • Allows the user to perform geographic analysis based on performing comparative analysis of scenarios intersection, overlapping of different objects, etc. • The amount of output information is too high for decision-making Environmental damage • Allows for integration of many types of emissions into a • No common and recognised measurement unit of calculation methodology, single measure of environmental damage environmental damage Russia (1983, 1999) • Explicitly takes into account geographical peculiarities of • No account taken of the receptors of polluting the given territories emissions Delphi method • Allows the user to analyse complex situations with • Subjectivism of estimates uncertain information and/or lack of time/resources for • Possibilities of unequal understanding the decision-making using experts problem in question by the experts Environmental impact • Allows detailed examination of all the impacts from • Very expensive in terms of time, resources, data assessment (EIA) specific sites and technologies demands • Can combine economic, environmental and social • Necessary to combine with dispersion modelling information • Very superficial types of studies • Focus is on the impacts and not the waste system itself Pollution dispersion models • Show detailed spatial distribution of emissions given the • Substantial computational power is needed (esp. relief, climate and the characteristics of the source of for multiple sources) emissions • A very expensive tool • Difficult to analyse the impacts on the final recipients In summary, we can say that LCI is good at modelling the were combined into a geo-spatial EIA (Patil et al., 2002; waste system but is only a first stage in identifying environ- Antunes et al., 2001), GIS and MCDA were combined by Dai mental impacts as it concentrates on emissions to water, et al. (2001), LCI and MCDA were integrated by Powell et al. land and air, but does not provide any indication of the (1996), and Munda and Romo (2001) and Powell et al. (1999) impact or significance of emissions locally. It needs to be integrated a simple multi-criteria approach to examine envir- integrated with other techniques such as EIA which can pro- onmental impacts from alternative waste management sce- vide the impact analysis needed based on the siting of infra- narios for the city of Bristol. structure, or movement of waste and with some optimisation In summary, these studies are still limited and cannot be procedure that can begin to deal with the issues of trading off used to solve regional waste problems because they have not economic costs and benefits against the social and environ- elaborated all the complex of factors influencing waste man- mental impacts of alternative waste management systems. agement processes at the regional level—namely spatial dis- Thus, it is clear that what is required is a combination of tribution of waste arisings, impacts of transportation and several methods in order to perform the complex analysis processing of waste as well as multidimensional character of the potential development of the municipal solid waste of these emissions, time dimension of waste generation, system. building new or expanding existing facilities and spatial dis- Several studies have already tried to combine some of tribution of impacts of waste treatment processes on these methods. During the 1990s, for example, GIS and EIA humans and valuable ecosystems. This paper reports on
  6. 6. 120 EC O LO GIC A L E CO N O M ICS 5 9 ( 2 00 6 ) 1 1 5 –1 30 Waste treatment facilities Incineration with energy recovery Civic amenity sites Landfills Materials recycling facilities Scrapyards Transfer stations Administrative borders Administrative borders of wards Fig. 2 – Location of the waste treatment plants in Gloucestershire. research (Shmelev, 2003; Shmelev and Powell, 2004) that has different methods a more useful tool might be developed for focused on the integration of three different approaches to the development of strategic municipal solid waste manage- the spatio-temporal analysis of the MSWM problem, namely ment plans. The aim therefore was to develop an integrated a life cycle inventory approach (LCI module), which helps to technique that would give useful support to the decision- identify emission patterns within the MSWMS, a multi-cri- maker regarding the potential development paths and trade- teria optimisation approach (MO module), which helps to offs between economic and environmental performance of find compromise solutions among environmentally, eco- alternative waste systems. nomically and socially preferred options, and a geographic Research carried out in Russia (St. Petersburg and the information systems approach (GIS module), which provides region) and the UK (Gloucestershire) has concentrated on a a base for siting waste management facilities, transpor- complex analysis of the MSWMS, taking ecological, economic tation, social impacts, as well as locating environmental as well as social aspects of the management of municipal impacts on valuable ecosystems. A Russian approach to solid waste into account. calculating environmental damage was utilised to weight Due to software limitations, it was decided to limit the the importance of different sub-territories covered by the analysis of the municipal solid waste management system system. It is hoped that this approach will provide a new to examination of five major components: i.e. economic perspective for the analysis of municipal solid waste man- costs of running the system, public health, the state of agement systems. the flora and fauna, saving of material resources and land- scape quality. Four out of the five factors chosen to char- acterise the waste management system relate to the main 5. Development of the integrated methodology goals of the EU Landfill Directive (European Council, 1999) (reduction of adverse effects of the landfill of waste on the Based on an understanding of the weaknesses of the methods environment, in particular on surface water, groundwater, identified in Table 1 above, it was decided that by combining soil, air and human health) and also correspond to the
  7. 7. EC O L O G IC A L E C O N O M IC S 5 9 ( 2 0 06 ) 11 5 –1 30 121 most relevant subject themes of the UN Sustainable Devel- opment Indicators: social (health), environmental (atmo- 6. Description of the modules within the sphere, biodiversity) and economic (consumption and integrated method production patterns) (United Nations, 2001). The fifth com- ponent, ‘landscape quality’, was selected to reflect the 6.1. The GIS module important role that landscape plays in local communities as stated in the European Landscape Convention, which The key elements of the GIS module are the digitised maps of identifies landscape as “a key element of individual and the county of Gloucestershire, UK, obtained from a range of social well-being”. Under Chapter II of the Convention, different sources. The maps are overlaid and allow graphical signatories agree to “integrate landscape into…regional analysis of the location of the physical waste infrastructure, and town planning policies…as well as in any other and transport routes in relation to the environmentally sensi- policies with possible direct or indirect impact on land- tive areas and the centres of population density. The census scape” (Council of Europe, 2000). These five components ward was taken as a minimal geographical unit for population offered a relatively simple and straightforward means of data. analysing economic–environmental trade-offs. The mone- tary costs of operating a waste management system are 6.2. The impact assessment module of critical concern to local authorities, materials savings (i.e. recycling) are of national concern, and for strategic It should be noted that the methods of the analysis and com- decision-making purposes both these elements need to parison of the emission inventory results within life cycle be directly compared to the impacts on environmental analysis is an area open to debate. In some cases, the list of and social factors (i.e. human health, environmental the emissions analysed numbers several hundreds items. In ‘health’ and landscape quality). The data on the selected order to deal with this vast amount of information in the components were also available and relatively easy to current research, the methodology expressed in the Vremen- obtain. naja metodika (1999) and Vremennaja tipovaja metodika Life cycle analysis using the Proctor and Gamble (2001) (1983) was taken as an instrument for comparing scenarios model was integrated with a GIS and an optimisation techni- with heterogeneous outputs. The list of substances taken into que. The LCA model allowed the researchers to examine a account in the analysis can be seen in Appendix A. The toxi- wide range of emissions from alternative waste management city coefficients database for all the pollutants allows conver- scenarios; the GIS allowed actual and proposed waste man- sion of the wide spectrum of the different substances into a agement sites, along with ecologically sensitivity of the land- unified index of the environmental damage, which lessens scape to be mapped; the single criteria optimisation technique the dimension of the problem substantially and simplifies permits the possibility of deriving a unique solution of the the solution procedure.2 problem. The method is used here to provide the spatial dimen- The most difficult choice was that of the optimisation sion of environmental damage around waste treatment procedure. There are well known, fast and reliable methods infrastructure sites in the form of the coefficients of the for linear problems, whereas it gets more complex when importance (significance) of the territories around the the situation requires mixed integer programming. There waste treatment plants. Such coefficients were derived by are ways for reducing the multidimensional problems to performing a series of operations on the GIS maps. The single-criterion ones, however, and taking account of sev- dispersion of pollutants from the various waste treatment eral objectives simultaneously. The type of optimisation facilities was approximated by a 5-km radius circle around problem employed here (linear mixed integer programming each of the sites. The coefficients of significance were problem) is complex and demands significant computa- derived based on the weighting of sensitive areas by a tional power and efficient algorithms, especially for the group of experts based at the University of Gloucestershire real scale modelling. Constraints on resources and compu- using a Delphi approach. Standard national designations of tational power led to a focus on a two-dimensional pro- ecological and landscape importance were utilised by the blem by examining single-criteria overall system cost experts: Sites of Special Scientific Interest (SSSI), National minimisation with simultaneous calculation of an addi- Nature Reserves (NNR), Special Areas of Conservation (SAC), tional parameter (such as the environmental damage Specially Protected Areas (SPA), RAMSAR sites and an indi- caused by the system performance). Although limited this two-dimensional solution space still provides a useful 2 The approach used here was one developed by the Russian starting point for understanding how useful such an inte- environmental economic school, which is explained in Vremen- grated methodology might be. naja metodika (1999) and Vremennaja tipovaja metodika (1983). The large sets of heterogeneous data used in the model The main idea of calculating environmental damage according to (geographical, economic, environmental and social) are inte- the Russian methodology consists of integrating the amounts of grated using relational database technology. The database emitted pollutants into a single index of environmental damage. The actual emissions of polluting substances are multiplied by system consists of several interrelated tables representing coefficients of environmental harm, which are in the inverse different aspects of the problem under study (e.g. different relation to the MAC (maximum allowable concentrations) of types of waste analysed, spatially distributed waste genera- pollutants in question. MAC are based on the results of tion centres, a range of waste treatment facilities, multitude substantial medical and environmental research (i.e. toxicological of emission types, etc.). data). The list of coefficients can be found in Appendix A.
  8. 8. 122 EC O LO GIC A L E CO N O M ICS 5 9 ( 2 00 6 ) 1 1 5 –1 30 cator of population density. The average number of people For each of the types of waste mentioned three basic living within 5-km circle around the waste treatment plants treatment technologies are analysed: recycling, incineration was calculated using the average population density of with energy recovery and landfilling. The emissions to air, neighbouring wards covered by the circle. The experts water and soil are analysed. In order to get the integral index were asked to rate their perceptions of the relative sensitiv- of environmental damage, the amounts of the polluting ity of the designated areas to the potential emissions from emissions are multiplied by the respected coefficients of the waste management facilities on a 1–10 scale. This infor- environmental harm, according to Vremennaja metodika mation was then integrated into a Randomised Preferences (1999) and Vremennaja tipovaja metodika (1983) as described method (Hovanov, 1996), which took the relational data above. derived from the experts to derive weighting factors. A Del- phi approach was utilised because of the complexity and 6.4. Optimisation module uncertainty over the impacts of the regional waste manage- ment system on the different aspects of the human and This module integrates the information on plant locations natural environments. The Delphi approach provided a and distances between the centres of population density and quicker and cheaper alternative to more narrowly defined waste treatment plants from the GIS module and the informa- pollution dispersion modelling approaches. The Delphi tion on the amounts of emissions from each type of waste, approach has the added advantage of enabling localised collected, sorted and treated by each of the technologies from priorities to be integrated into the significance measures, the LCI module. It is here that the choice of collection systems, but the methodology employed needs to be transparent in sorting and treatment technologies, as well as geographical order to understand the trade-offs generated. Thus, such an distribution of the waste management facilities are optimised approach will provide variations in significance measures in over the time period of interest according to the total system different regions, related to population geography and pro- cost minimisation criteria. The problem that is being analysed tected areas. here belongs to the class of linear mixed integer programming Then overall indices of the importance of the territories problems. LINGO optimisation software uses branch and around the waste treatment plants were obtained. First, all bound methods to solve problems of this type. The informa- the 5-km radius circles around the waste treatment plants tion on the problem dimensions for the Gloucestershire case were overlayed onto the maps of the different types of study is laid out in Table 2. sensitive areas (including centres of population density). The initial problem is set in a single-criteria cost mini- Then the percentage of the intersection of each circle by misation framework. The reason for this is that all the each type of the sensitive area was multiplied by the impor- improvements in the environmental performance of the tance factor for a given sensitive territory and it was sum- waste management systems are bound by the budgets of mated over all six types of areas analysed according to the the related administrative units, and cost minimisation is formula: still the dominant criteria for waste management system development. The environmental damage is calculated X J SðCk Ej Þ here as a by-product of the minimum costs scenario accord- Ik ¼ Fj T j¼1 SðCk Þ ing to the formula: X K X K where Ik is the importance score of the circle around kth ED ¼ Ik T gI TElk ; waste treatment plant; Fj is an importance factor for the k¼1 k¼1 environmental sensitive territory type j, j = 1…J; S denotes where ED is a total systems environmental damage, Ik is an area; C is an area of the circle around a waste treatment importance score of the territory around kth waste treat- plant; Ej is a joint object consisting of the parts of environ- ment plant, Elk is the amount of emissions of the lth type mentally sensitive areas falling within a given circle Ck. (l = 1…L) (from the LCI module) and γl is an environmental The borders of each of the geographical objects are stored damage coefficient for the emission type l. in the digital database with additional information such as The final two-dimensional solution space in the form pre- the name of the object, areas and geographical coordinates. sented in Fig. 7 is obtained by performing a sensitivity The centre points of the census wards are used to define the waste generation places, and transport routes are considered here as the links connecting the centroids of the wards and Table 2 – Gloucestershire real problem dimensions the waste treatment plants. Set Definition Quantity of elements 6.3. The LCI module J Waste generation points 145 H Waste types 9 I Waste treatment centres 86 In the framework of the analysis carried out here, the life cycle K Waste treatment technologies 6 analysis is bounded on the one side by the post-consumption T Periods of system functioning 20 generation of waste and on the other side by final disposal. It includes the analysis of the municipal solid waste stream Which amounts to 13,560,480 variables in the mixed integer pro- comprising eight components—paper, glass, ferrous and gramming model, including integer variables: 92,880, real variables: 13,467,600 and number of constraints: 13,591,278 (without trivial non-ferrous metals, plastics (film), plastics (rigid), textile, constraints, stating non-negativity of decision variables—123,678). organic and “other”.
  9. 9. EC O L O G IC A L E C O N O M IC S 5 9 ( 2 0 06 ) 11 5 –1 30 123 analysis, which changes the waste treatment capacities, in Table 3 – Composition of municipal solid waste in relation to landfill space. It allows the decision-maker to ana- Gloucestershire, 1998/1999 lyse the given waste management system in terms of the Material Gloucestershire National trade-offs between environmental and economic objectives.3 average, % average, % The simplified version of the model utilised here based on Fines 2.3 7 the work of Baetz and Neebe (1994) was built using the LINGO Ferrous 4.0 6 7.0 optimisation software. The model only permitted exami- Glass 3.4 9 nation of a reduced set of problem dimensions: three waste Green 11.3 21 treatment technologies—recycling, incineration and landfill- Putrescibles 34.4 ing were considered over 10 time periods with no considera- Misc. com. 5.8 8 Misc. non-com. 0.5 2.20 tion of space dimension. The problem's dimensions were Non-ferrous 1.0 2 determined by the constraints in the number of boolean and Paper and card 20.8 32 continuous variables in the Demo version of LINGO 7.0. Due to Plastic film 4.9 5 the resource constraints (software limitations), the model has Rigid plastics 7.6 6 not been realised to its full potential. Textiles 3.9 2 Two versions of the model were developed, which use Gloucestershire figures do not include the recycled waste. Open Data Base Connectivity (ODBC) technology for transfer- ring data from one software package to the other. The useful- ness of the developed model is that it allows the user to 6% in Gloucester to 19% in Cotswolds District. The dominant change the initial data outside the model and then to “plug- municipal solid waste treatment method is landfilling (82% in in” the new datasets for subsequent solving. It is useful in the the South West Region of the UK). case of sensitivity analysis involving large number of para- The average composition of municipal solid waste in Glou- meter changes. cestershire is presented in Table 3. It should be stressed here that solving this problem in real Fig. 2 illustrates location of waste facilities and Table 3 dimensions with standard tools would require handling vec- compares waste composition to the UK national average. tors of model variables with 15,000,000 components. This will Simulations using the dataset for Gloucestershire were definitely require using more powerful database management performed on the simplified version of the model. 8140 systems (e.g. SQL Server) and the problem could be solved simulations were undertaken (see Fig. 7), where the waste faster if its special structure could be taken into account. treatment capacities for recycling, incineration and landfill- The problems that can be analysed include the choice of ing were changed. This could illustrate, for example, the waste processing technology (e.g. among landfilling, compost- possible consequences of introducing the EU Landfill Direc- ing, waste-to-energy incineration, recycling), waste manage- tive in the county, which may result in fewer landfills and ment facility plant siting and optimisation of the whole increased recycling capacity, with consequent impacts on MSWMS performance using different goal functions. transport routes and costs across the county. The objective The next section examines a simple application of the of the directive is to prevent or reduce as far as possible model to waste management in Gloucestershire. negative effects on the environment from the landfilling of waste, by introducing stringent technical requirements for waste and landfills, and reducing the quantity of biodegrad- 7. Case study of Gloucestershire able material going to landfill. The scenarios examined here show the potential effects of reductions in available landfill Gloucestershire lies in the west of England (South West space as a result of the Directive and explore the impacts Region), has a total area of 2,618,000 km2 and a population of of increased tipping fees and recycling subsidies on the 574,000 (2001). Gloucestershire comprises six local authorities: environmental and economic performance of the system. Cheltenham Borough, Cotswolds District, Gloucester City, For- Fig. 7 illustrates the combinations of minimal costs and est of Dean District, Stroud District and Tewksbury Borough. corresponding environmental damages for the whole Average number of people in the households is 2.41. The range of scenarios examined. All the combinations of average disposable income per person per year is £10,073 potential environmental damage and economic costs are (1999, data for the South West Region), and annual waste given here under equal economic conditions. Only the arisings range from a low of 280 to a high of 432 kg of muni- landfill and waste treatment capacities were changed in cipal solid waste per person per year is produced in Glouces- this analysis. tershire. The annual recycling rate in 1998/1999 ranged from 8. The description of the results of the 3 Description of the software used—LINGO, ACCESS, MAPINFO, simulations experiments Procter and Gamble LCI MODEL.The integration of the hetero- geneous software was necessary for building the working The results of a series of simulation experiments are depicted interactive modelling system. In the current research, the in Figs. 3–7. The study of the developed model of the regional optimisation software package LINGO 7.0 (demo version), GIS package MapInfo Professional 6.0, Database management system waste management system was conducted along the follow- MS Access 2000 and spreadsheet MS Excel 2000 were used, along ing main lines: it was decided to study the sensitivity of the with the Procter and Gamble LCI model. model first to the changes in technological parameters of the
  10. 10. 124 EC O LO GIC A L E CO N O M ICS 5 9 ( 2 00 6 ) 1 1 5 –1 30 Total costs and environmental damage ED 5800000 700000000 C 5700000 600000000 Total system 5600000 management 500000000 costs, GBP 5500000 400000000 5400000 300000000 Index of 5300000 environmental 200000000 damage 5200000 5100000 100000000 5000000 0 10000 1000 2000 3000 4000 5000 6000 7000 8000 9000 L Fig. 3 – Scenario 3. RE = 200, W = 200, LL = 5000. available capacity of the existing landfill under different com- tutional transition to recycling part of the waste takes place, binations with other technological parameters being fixed, causing considerable decrease in environmental damage by a and later to the changes in price parameters—the cost of factor of 1.08. recycling of a ton of waste and costs of collection and trans- In scenario 2, there is an option of opening a small addi- porting waste to the landfill. The combinations of the system tional landfill with the capacity of 1,000,000m3. The first local parameters used in sensitivity analysis are shown in Table 4. minimum of environmental damage is found at L = 5500. Such ED—environmental damage denotes the index of environ- a sharp decrease in environmental damage is caused by the mental damage and C—costs denotes total management growth in recycling, instead of harmful landfilling in the costs in British pounds. landfill; the following growth in damage is caused by opening In the first scenario, the recycling and incineration capa- of an additional landfill in the 9th period; and the rapid cities were limited by 200,000 t/year, there is no opportunity to decrease in environmental damage starting at L = 4500 can open an additional landfill. With parameter L decreasing at be explained by the ever increasing rate of recycling. The first against a background of considerable growth in costs, the costs at the same time are starting to grow at a naturally slow growth in environmental damage takes place, caused by faster rate. intensive use of incineration as an alternative for the decreas- In scenario 3 (Fig. 3), the rapid decrease in environmen- ing landfill capacity; then with decreasing L b 5500 the substi- tal damage as L approaches the value of 5500 is caused by ED Total costs and environmental damage C 7000000 900000000 800000000 6000000 700000000 5000000 Total system 600000000 management costs, GBP 4000000 500000000 3000000 400000000 Index of 300000000 environmental 2000000 damage 200000000 1000000 100000000 0 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 L Fig. 4 – Scenario 4. RE = 600, W = 200, LL = 0.
  11. 11. EC O L O G IC A L E C O N O M IC S 5 9 ( 2 0 06 ) 11 5 –1 30 125 Total costs and environmental damage ED 6000000 700000000 C 5000000 600000000 Total 500000000 4000000 management costs, GBP 400000000 3000000 300000000 Index of environmental 2000000 damage 200000000 1000000 100000000 0 0 55 65 75 85 95 105 115 125 135 145 A Fig. 5 – Scenario 6. RE = 600, W = 200, L = 5000, LL = 0, illustrating changes in A. the growth of the share of recycling; the rapid growth in placing the incineration residue becomes critical, the shift environmental damage after L b 4500 is caused by the open- towards recycling at a larger scale takes place. ing of a new landfill for 5,000,000 m3 in the 6th period In the 5th scenario, everything develops similarly to the with simultaneous decrease in the share of the recycling 4th; however, due to the larger planned incineration capacity and incineration. After that, the growth in the share of and smaller recycling capacity, the shift to the second stage of waste being incinerated is growing. The following local intensive recycling takes place later, at about L = 750, and to minima can be explained by the shifts of the moment the third–earlier–around L = 200. of opening an additional landfill to 5th, 4th period and so The sensitivity of the solution to the problem to the on. changes in price parameters is illustrated in Figs. 5 and 6. The tendency for the environmental damage and mini- Analysing the changes in environmental damage, caused mal management costs to change in scenario 4 (Fig. 4) could by the decreasing price of complex recycling of a ton of be divided into three different stages—10,000 N L N 5500, waste (parameter A, recycling costs, Fig. 5), we come to 5500 N L N 100, L b 100. In the first stage, the gradual growth the conclusion about the lack of changes in environmen- of the share of the waste being incinerated takes place, tal damage with parameter A being reduced from 145 to which causes the slow growth in environmental damage 110. Then the sharp decrease in environmental damage— and costs; decreasing environmental damage and costs more than by a factor of 1.7 with the following decrease growing at the faster rate in the second stage are caused in A to 80, and again, at the interval [55…80] environ- by the growth in the share of waste undergoing complex mental damage is at the lower than in the first case, but recycling at L b 100; when the landfilling capacity even for stable level. Total costs and environmental damage ED 6000000 800000000 C 700000000 5000000 600000000 Total management 4000000 costs, GBP. 500000000 3000000 400000000 Index of 300000000 environmental 2000000 damage 200000000 1000000 100000000 0 0 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 10 20 30 40 50 60 70 80 90 B Fig. 6 – Scenario 7. RE = 600, W = 200, L = 5000, LL = 0, illustrating changes in B.
  12. 12. 126 EC O LO GIC A L E CO N O M ICS 5 9 ( 2 00 6 ) 1 1 5 –1 30 6,000,000 5,500,000 Indicator of potential environmental damage 5,000,000 4,500,000 4,000,000 3,500,000 3,000,000 2,500,000 2,000,000 1,500,000 400,000 450,000 500,000 550,000 600,000 650,000 700,000 750,000 800,000 Total management costs, thous. GBP Fig. 7 – Two-dimensional solution space. Changes in the parameter B—costs of collection and The shape of the thick curve representing the set of transportation of waste to the place of their landfilling in non-dominated solutions (solutions that are equal or not the Landfill 1 could suggest the optimal level for the trans- worse off than all the others) depicts the peculiarities of port costs set up in the interests of the environmental pro- the complex problem of the development of a waste man- tection (under conditions of legal waste discharges by the agement system giving the decision-maker the range of companies and municipalities). The results of the simula- options he or she could choose from and thereby helping tions experiments that could be seen in Fig. 6 show that, him trade-off economic versus environmental aspects of with the transport costs being increased up to a certain level the development of the system in question. We are deci- (in our case B = 120) and given the laws are observed, trans- sively not proposing the decision-maker “the best solu- porting waste to the landfill may become less desirable than tion” or BPEO, but providing him or her a freedom of recycling. informed choice; however, hard it may be to make one. The main result of the work—two-dimensional solution The latter appears in the realm of pure political decision- space, which is a integration of the results of sets of simula- making. tion experiments 1 to 5 (Table 4), shows that, by increasing total system management costs by a factor of 1.82, it is possi- ble to diminish the total environmental damage by a factor of 9. Discussion 2.99. The results presented here illustrate an application of a sim- plified ecological–economic model of a municipal solid waste Table 4 – Parameters changed in sensitivity analysis management system. Full development of the model would Set of LL RE W L Changed The allow solution of more complex problems involving real deci- simulation parameter interval of sions of siting, choice of treatment technology, collection and experiments change sorting method. Certain weaknesses remain in the approach 1 0 200 200 L 10:10,000 taken here, primarily software limitations and probably lack 2 1000 200 200 L 10:10,000 of pollution dispersion modelling. 3 5000 200 200 L 10:10,000 The main strength of the model is that it allows the deci- 4 0 600 200 L 10:10,000 sion-maker to analyse the ecological–economic trade-offs in 5 0 400 400 L 10:10,000 the development of the municipal solid waste management 6 0 600 200 5000 A 55:145 system. It examines possible strategies of the development of 7 0 600 200 5000 B 10:300 the system, taking into account different siting options, Key: L—available capacity of the existing landfill, thousands m3 per choice of waste treatment technologies, performs preliminary year; LL—available capacity of the additional landfill, thousands investment planning and explicitly takes account of spatial m3 per year; RE—recycling capacity, thousand of tons per year; W— dimension of environmental impacts on public health and incinerating capacity, thousand of tons per year. valuable ecosystems.
  13. 13. EC O L O G IC A L E C O N O M IC S 5 9 ( 2 0 06 ) 11 5 –1 30 127 In the life cycle analysis performed here, the boundaries waste and work on material flows accounting of products are defined by the post-consumption solid waste generation entering the system in the first place, then with programming through to the moment of it's final disposal. If the improvement a full scale decision support tool for strategic boundaries were altered to include elements related to regional waste management could be created. The next step is the production of the waste processing equipment, trans- to apply more powerful software, possibly integrate pollution portation fuel life cycle, analysis of materials and pro- dispersion models for all sources of pollution and analyse ducts the solid waste was derived from, the results could more rigorously the chains of impacts. It could be valuable change significantly. to integrate the analysis of environmental impacts of trans- The model presented in this paper could be developed portation, take into account noise and congestion impacts. further to take into account the real dimensions of the pro- Models of this type could then be expanded and applied at blem, such as transportation of waste, improved pollution the regional level in the EU, to provide improved information dispersion models and introduce hyperbolic discounting on the tradeoffs to be made what are inherently difficult (Daly and Farley, 2004). If we take into account the origins of political problems. Appendix A. The list of emission coefficients Sector of the ecosystem Emission type Recycling Incineration Landfilling Damage coefficients Air Particulates 0.00327 0.00002 0 2.7 Air CO 0.00228 0.0004 3.125E−06 0.4 Air CO2 0 1.1293 0.2209825 0.4 Air CH4 0 0 0.098215 0.7 Air NOx 0.00231 0.0016 0 16.5 Air N 2O 0.000053 0 0 30 Air SOx 0.003947 0.0003 0 20 Air HCl 0.0000033 0.0001 1.625E−05 20 Air HF 5E−09 0 3.25E−06 500 Air H 2S 0.000012 0 0.00005 500 Air HC 0.001692 0.0001 0.0005 20 Air Chlor. HC 0 0.0001 8.75E−06 50 Air Dioxins/furans 0 5E−13 0 50,000 Air NH3 0.0000004 0 0 28.5 Air As 0 0.0000025 0 500 Air Cd 0 0.0000005 1.4E−09 500 Air Cr 0 0.0000063 1.65E−10 1670 Air Cu 0 0.0000063 0 500 Air Pb 0 0.0000063 1.275E−09 5000 Air Hg 3E−09 0.0000005 1.025E−11 5000 Air Ni 0 0.0000025 0 500 Air Zn 0 0.0000063 1.875E−08 500 Air Landfill gas (250 nm3/t) generation (t/t) 0 0 250 0 Water BOD 0.00239 0 0.0004751 5 Water COD 0.02084 0 0.0004751 2 Water Sus. sol. 0 0 0.000015 0.15 Water TOC 0.000004 0 0.0000003 50 Water AOX 0.0000025 0 0.0000003 1000 Water Chlor. HCs 0 0 1.545E−07 0 Water Dioxins/furans 0 0 4.8E−14 0 Water Phenol 0 0 5.7E−08 0 Water NH4 4.47E−07 0 0.0000315 1 Water Tot. metals 0 0 1.442E−05 0 Water As 0 0 2.1E−09 90 Water Cd 0 0 2.1E−09 250 Water Cr 0 0 9E−09 550 Water Cu 0 0 8.1E−09 550 Water Fe 0 0 1.425E−05 1 Water Pb 0 0 9.45E−09 11 Water Hg 0 0 9E−11 15,000 Water Ni 0 0 2.55E−08 90 Water Zn 0 0 1.02E−07 90 Water Cl 0.000011 0 0.0000885 550 Water F 9.7E−07 0 5.85E−08 550 Water NO3 0 0 0 0.2 Water S- 0.000006 0 0 550