You are on page 1of 17

Waste Management 27 (2007) S58S74 www.elsevier.

com/locate/wasman

Landll modelling in LCA A contribution based on empirical data


Gudrun Obersteiner *, Erwin Binner, Peter Mostbauer, Stefan Salhofer
Institute of Waste Management, Department Water Atmosphere Environment, BOKU University of Natural Resources and Applied Life Science, Muthgasse 107, 1190 Vienna, Austria Accepted 16 February 2007 Available online 11 April 2007

Abstract Landlls at various stages of development, depending on their age and location, can be found throughout Europe. The type of facilities goes from uncontrolled dumpsites to highly engineered facilities with leachate and gas management. In addition, some landlls are designed to receive untreated waste, while others can receive incineration residues (MSWI) or residues after mechanical biological treatment (MBT). Dimension, type and duration of the emissions from landlls depend on the quality of the disposed waste, the technical design, and the location of the landll. Environmental impacts are produced by the leachate (heavy metals, organic loading), emissions into the air (CH4, hydrocarbons, halogenated hydrocarbons) and from the energy or fuel requirements for the operation of the landll (SO2 and NOx from the production of electricity from fossil fuels). To include landlling in an life-cycle assessment (LCA) approach entails several methodological questions (multi-input process, sitespecic inuence, time dependency). Additionally, no experiences are available with regard to mid-term behaviour (decades) for the relatively new types of landll (MBT landll, landll for residues from MSWI). The present paper focuses on two main issues concerning modelling of landlls in LCA: Firstly, it is an acknowledged fact that emissions from landlls may prevail for a very long time, often thousands of years or longer. The choice of time frame in the LCA of landlling may therefore clearly aect the results. Secondly, the reliability of results obtained through a life-cycle assessment depends on the availability and quality of Life Cycle Inventory (LCI) data. Therefore the choice of the general approach, using multi-input inventory tool versus empirical results, may also inuence the results. In this paper the dierent approaches concerning time horizon and LCI will be introduced and discussed. In the application of empirical results, the presence of data gaps may limit the inclusion of several impact categories and therefore aect the results obtained by the study. For this reason, every eort has been made to provide high-quality empirical LCI data for landlls in Central Europe. 2007 Elsevier Ltd. All rights reserved.

1. Introduction In recent years waste managers, planners and local authorities have been faced with increasing demands to deliver a sustainable approach to waste management and to integrate strategies capable of producing the best practicable option for the environment. Waste management planning has become an institutionalised element in public

Corresponding author. Tel.: +43 1 3189900 318; fax: +43 1 3189900 350. E-mail address: gudrun.obersteiner@boku.ac.at (G. Obersteiner). 0956-053X/$ - see front matter 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.wasman.2007.02.018

planning eorts in all EU Member States. One important aspect of waste management plans is to ensure the identication of areas in which specic measures should be taken to reduce the environmental impacts of waste management. To demonstrate the performance of management alternatives in the decision-making process, authorities, communities, industry and waste management companies should consider environmental aspects in addition to the evaluation of technical and economic aspects. It is accepted that life-cycle assessment (LCA) concepts and techniques provide solid waste planners and decision makers with an excellent framework to evaluate MSW management strategies.

G. Obersteiner et al. / Waste Management 27 (2007) S58S74

S59

The rst LCA studies to include the whole life cycle of a product in environmental assessment were introduced in the 1990s. Life-cycle assessment was initially developed for the purpose of analysing products, although in the meantime it has also been applied to dierent services. One of these services is represented by the treatment of a particular amount of waste. However, it should be borne to mind that conventional LCAs were designed to assess products rather than services. In the meantime, waste management evolved to a separate section within LCA. Since 1998 the International Expert Group on Life Cycle Assessment for Integrated Waste Management has dealt with this topic (Coleman et al., 2003). Software tools have been created dealing with waste management issues alone (e.g., ORWARE (Eriksson et al., 2002), LCA-LAND (Nielsen and Hauschild, 1998), IWM2 (McDougall et al., 2001), WISARD (http:// www.ecobalance.com/uk_wisard.php)) or as a main issue like Ecoinvent 2000 (Frischknecht and Rebitzer, 2005). Initially, LCA experts dealt chiey with waste management problems and focused on methodological questions. Nowadays an increasing number of people working in the waste management sector use life-cycle assessment as methodology to support decision making. The rst group is clearly involved in methodological problems resulting from life-cycle assessment of waste management processes, whilst the second is interested more in specic waste management application. The use of LCA for resource and waste management issues implies a slightly dierent focus than traditional product-oriented LCAs. Most product LCAs do not consider end-of-life phases or assume a simplied form of disposal. Despite intensive research on this topic during recent years, the development of a specic modelling of the fate of the substances contained in waste disposed of through incineration, landll or recycling, is, in view of the current state of scientic knowledge, not an easy task. With particular regard to the landlling of waste or residues of municipal solid waste incineration (MSWI), respectively, or to mechanical biological treatment (MBT), two major challenges are posed by the use of LCA: data gaps and methodological decisions. Data pertaining to waste collection, recycling and treatment all processes where direct measurements are possible are, on the whole, more reliable than data from landlls which partially have to be modelled and where estimations are necessary. A long-range perspective makes experiments and eld studies on landlls dicult to perform and therefore the uncertainty with landll models is considerable. On top of this is the fact that over the last decade landll design has changed. In the future, according to national and EU legislation (landll directive) further improvements are expected. With more modern waste management strategies including up-to-date treatment technologies such as MBT or MSWI, new types of landll will have to be faced. At present due to dierent reasons, there is a lack of high quality data for these types of land-

ll. In addition to problems of data generation, methodological issues have to be solved, particularly with regard to the time factor where dierent ideas as to which period of time emissions from landll should be considered are under discussion (Finnveden, 1999). In the mentioned software models, landll is included. However, these landll models are based on a number of assumptions and predictions about future processes in a landll. Therefore,in some studies only the operational stage of waste treatment plants is considered. Several authors rightly insist on the importance of the burdens from landlls within the disposal process chain (Hellweg et al., 2003; Doka, 2003; Finnveden, 1999; Sundqvist, 1999). In practice there is still no panacea on how to deal with the problems emerging when modelling landlls within LCA of waste management strategies. The authors are aware that since the late 1990s in the LCA community, the mentioned problems for LCA in waste management have been under discussion and many authors have published their views on this topic (e.g., Sundqvist et al., 1997; Finnveden, 1999; Bjarnadottir et al., 2002; Doka, 2003). In this paper therefore the focus is set on providing waste managers with a commented summary of the ongoing discussion from a waste management point of view. In the past years dierent multi-input inventory tools have been developed, most of which are based on the results of empirical data. Especially in the case of future MBT landlls, there is a scarcity of sound data. Often landll emission data from the literature are used in a rather arbitrary way. Accordingly, the authors wish secondly to point out relevant points which should be taken into account when using the literature data and thirdly highquality empirical LCI data for landlls in Central Europe are provided. 2. Methodological considerations 2.1. LCA in waste management In talking about life-cycle assessment of integrated solid waste management, one can distinguish numerous strategic issues which may require dierent methodological solutions: 1. LCA of a specic product (whole product life cycle or just waste management system for a given material): the whole product life cycle as well as the waste treatment after product use is included within the system boundaries. 2. LCAs evaluating waste treatment options: assuming that all upstream impacts are equal, the life cycle of waste starts when products are disposed of in the trash bin and ends when the waste material is degraded or returned to the technological system through recycling to replace other products. The system boundaries can be set where the waste is introduced into the system.

S60

G. Obersteiner et al. / Waste Management 27 (2007) S58S74

3. LCAs to optimise waste management strategies: In many areas this case is similar to alternative 2 but the focus is placed on the eect of waste management strategies. For instance therefore waste prevention may be included, implying the need to enlarge the system boundaries. Therefore, an LCA focusing on waste is based on dierent premises with regard to system boundaries and data collection. 2.2. Choice of time frame When applying an LCA approach for waste management, major data gaps emerge for the long-term emissions from landlls, namely for emissions into water. The present level of knowledge does not allow secure statements to be made concerning the long-term amount and quality of leachate either for MBT or for MSWI landlls (Soyez, 2001). Existing emission monitoring systems for landlls are not designed to monitor emitted loads, but mostly reduced to periodic analysis of the leachate or groundwater composition. Waste in landlls will have a considerably long-lasting impact on the environment. Several authors including Sabbas et al. (1998), Sundqvist et al. (1997), Hellweg (2000), and Doka and Hischier (2005) conclude that the pollutant potential remaining in a landll 100 years after waste placement is signicant. In general the environmental potential of landlls depends largely on the specic material or product and the type of landll under investigation. Heavy metals for example tend to concentrate in landlls and are washed out to a varying degree over time. Unger (2005) identied in this context three major groups of materials:  materials with a lower environmental impact in the extraction and processing stage than in the use and disposal stage;  materials with a higher environmental impact in the extraction and processing stage than in the use and disposal stage; and  materials with an equal distribution of their environmental impact across the whole life cycle. Nevertheless long-term emissions may represent an important burden from landlls. Doka (2004) showed that the relevance of the disposal (including incineration and/or landlling) may reach 18% of the entire burden from production and disposal in the case of PVC. On the other hand, especially for the newer types of landlls (MBT-landll or MSWI landll), in the case of assessment of entire waste management systems (including transport, collection recycling (incl. credits), treatment and landll), the relevance of landlls to the overall results is decreasing remarkably (Wassermann et al., 2005). During the other stages in the lifetime of a product, emissions will occur more or less instantaneously or at least

within a limited time period. Emissions from other treatment methods such as incineration and composting also cause emissions that occur more or less immediately. Landll experts agree that landlls cannot be regarded as stable systems as to long-term emissions (e.g., Lechner, 2001; Sabbas et al., 1998). The designated barrier systems inertisation, solidication, sealing sheets etc. deteriorate in time and have a limited functional lifetime. There is a relevant and plausible potential that the remaining pollutant load in a landll will be released completely if long enough time spans are considered (Doka, 2003). Over the years in general three dierent phases for landlls can be distinguished (Sabbas et al., 1999). The shorttime phase is characterised by the emplacement of waste and active and passive maintenance. The time horizon amounts to decades. In the medium-term phase there is no more active maintenance. External factors, environmental impacts on the landll body are more or less constant. The time horizon comes to centuries. These two phases can be concluded as foreseeable. In the long-term phase, where the time horizon reaches 104105 years, the external factors change. Developments are not foreseeable in detail. The challenge to be dealt with is to select an appropriate time interval and the time dependent emission function to be integrated over the selected time interval. In order to make the potential emissions from landlling comparable to other emissions during the life cycle, potential emissions have to be integrated over a certain time period (Finnveden, 1999). When time frames are discussed, there is a consensus that the emissions should be integrated over a so-called foreseeable period. However, the time frame for the foreseeable period may vary from 15 years to 50,000 or even 100,000 years in dierent studies (Sundqvist, 2002). The Society of Environmental Toxicology and Chemistry (SETAC) recommends that the emissions should be integrated over an innite time period; if this is not possible, a time interval of 100 years should be applied. In the Ecoinvent 2000 data base (Frischknecht et al., 2004), long-term emissions are dened as emissions occurring 100 years after present. They are reported in separate emission categories. The landll inventories include long-term emissions with a time horizon of 60,000 years after present (Doka, 2003). The quantication of the long-term (>100 years) impacts from landlling is a problem that has not yet been solved. In the absence of knowledge, some LCAs assume that these emissions are zero, appearing as a better choice than other treatment and disposal options. The toxicity impacts are also frequently neglected because of insucient science-based knowledge, making the assessments of hazardous waste treatment options dicult. On this account several studies have located the waste disposal outside the system boundaries, reporting only the amount of waste leaving the system boundaries but making no mention of the quality of the waste (Koblmuller et al., 2004; Bjarnadottir et al., 2002). An extremely smart solution for the time problem was identied by Melloni et al.

G. Obersteiner et al. / Waste Management 27 (2007) S58S74

S61

(2003) who dened their functional unit as 1 ton of waste sent to the landll and lying in place for 30 years. The selection of time interval is however, amongst others, an ethical question based on the fact that by limiting the exposure time, eects on future generations will be avoided. Landlls postpone the release of emissions from todays wastes into the future (Doka, 2005); therefore discounting future burdens in LCA studies introduces the risk of burden shifting for the future (Hellweg et al., 2003). When taking into account the general developments occurring over the last 100 years concerning, e.g., technical innovations, the prediction of landll development over exceptionally long time spans is vague at best. Beside further possible changes of the waste input itself, it depends on many uncertain parameters such as the development of geochemical weathering, climate conditions or vegetation (Doka, 2003). The concept of time frames leads to the necessity of not only describing the outputs but also the impact as a function of time. It is however impossible to ascertain the future characterisation values which must be applied to emissions in order to obtain the impact potential (Van der Ven, 1997). Thus, when establishing the proper time horizon for the analysis of environmental burdens from landlls, ethical issues strongly dictate the demand for adequate data which stand up to scientic plausibility tests. It would not however always appear to be necessary to give much thought to the time frame of an investigation. First, the matter of system boundaries must be settled. As to issues such as comparison of recycling, incineration and landlling as alternatives, the choice of the time horizon may represent a most important factor in providing results for other questions like treatment of municipal solid waste in MBT, incineration or direct landlling as alternatives those long-term emissions no longer have any eect. The latter is due to the fact that for dierent landll types the amount of pollutants released to water and air diers in time. These dierences however only play a signicant role in the rst 100 years. In sanitary landlls after the methane phase, pollutants continue to be released via leachate, although usually on a lower level. Emissions to air are negligible. After 100 years the leachate concentrations are assumed to decrease to the level of bottom ash landlls (Doka, 2003). There are, of course, distinct dierences between the chemistry in a bottom ash landll and a 100-year old sanitary landll, for example the content of organic carbon. Nevertheless, in absence of better data, e.g., the Econinvent database, bottom ash landll concentrations are considered representative once the municipal waste has stabilised to a certain extent and is no longer biologically reactive (Doka, 2003). The same assumptions can be made for MBT landlls: on comparing the leachate concentration of dierent landll types in the rst years and 30 years after landll construction one can nd hints referring to this. For example, during the rst years considerable dierences are revealed in concentrations of pH-value and COD (chemical

oxygen demand), as well as for heavy metals. These dierences decrease in time. It can therefore be concluded that if dierent waste management strategies resulting in dierent landll types are compared for long-term emissions (over 100 years), the amount of waste remaining in the landll is the only crucial factor pertinent to discrepancies in environmental pollution. 2.3. Data source: modelling versus empirical results In general, two dierent approaches are applied in the modelling of waste disposal processes. Empirical results from measurements obtained from various landlls of dierent ages can be used wherever basic conditions are similar to the specic landll. Data are determined in a black-box principle with, e.g., municipal solid waste or incineration residues or mechanical biological pre-treated waste as input and emissions as output incapable of reecting changes in waste composition. Multi-input inventory tools on the other hand take into account the initial waste composition. The causal relation between the specic waste input and the resulting emissions must be calculated by some kind of model. Some of these emissions are directly dependent on the chemical composition of the product studied. It should however be kept in mind that some emissions are very much process-dependent and are therefore dicult to predict. Within the multi-input inventory tools dierent approaches can be used for modelling:  To consider the theoretical maximum load, one can dene a maximum emission potential caused by the total pollutant content in the waste.  To calculate the behaviour of waste in a waste-specic manner, the degradability of waste fractions is used and release factors are introduced to portray the re-precipitation of degraded material within the landll. Laboratory tests to determine the availability of substances embodied in the waste are used.  Another possibility is to carry out model calculation depending on landll specic parameters (e.g., transport model or geochemical model). Dierent release factors are calculated for each chemical element and are calibrated according to eld measurements. Uncertainties in chemical and biological interactions and the preferential ow of leachate through the landll body makes projections dicult but may also be considered. Only calibrated models should be applied. Fig. 1 provides a review of methodologies currently applied to estimate the environmental loads caused by landlls. The methodologies vary as to starting material, represented both by data from specic landlls, as well as data from landll simulation reactors (LSR). In some cases an attempt was made to alter the starting material, e.g., by evaluating weathering processes as in the beneath described MPWLP-test (Sabbas and Lechner, 2001).

S62

G. Obersteiner et al. / Waste Management 27 (2007) S58S74

Fig. 1. Approaches for estimation of loads; T = Time horizon.

In the following the dierent possibilities will be introduced in brief and their pros and cons discussed. It should be mentioned at the outset that in general the multi-input inventory tools are to be preferred because of the need to meet the goal of mass balance. The characteristic of waste as a mixture of various materials adds another element of complexity to the use of LCA. Due to the mixed and variable composition of waste, it may be dicult to determine which materials in waste cause a given emission (http:// waste.eionet.eu.int/lca). For example there should be no emissions of heavy metals from the landlling of untreated wood. If empirical results are used for calculation one should be aware that LCAs require the acquisition of signicant amounts of data and that the quality of data determines the utility of the nal LCA (Bjarnadottir et al., 2002). After studying several case studies and databases, Finnveden (1998) concludes that data gaps limit the inclusion of several impact categories or cause them to be less covered and therefore limit the types of conclusions that can be drawn from these studies. Human and eco-toxicological impact categories are characterised by severe data gaps due to the large number of possible pollutants in the waste or produced by waste treatment and to the lack of knowledge of the behaviour of all these pollutants. One means of simplifying the process of data generation is to use legal limit values as maximum possible output from the landll. This approach can be used at least for countries where appropriate legislation regulates the amount of pollutants permitted in ground water or air by establishing legal limit values. However, similar legal limit values do not exist for all pollutants and indeed, values are only relevant for the rst decades of the landll. During this time there is a presence of strong reactivity, enabling limits to be exceeded. The technical barrier system as well as the leachate treatment of the landll for

instance, is assumed to function properly so that legal values can be met. Studies performed to investigate the prediction of the mid- to long-term behaviour of landlls were carried out particularly during the last two decades. Many are based on results gained from laboratory experiments, landll simulation reactors (LSR) with an experimental volume between 0.1 and 1 m3 of waste. Calculating the liquid (leachate)solid-ratio (LS-ratio) enables the results to be related to full-scale landlls if well stabilised material is used. This methodology indirectly implies a similar moisture distribution in full-scale landlls and LSR (Doberl et al., 2003). Variations in the LS ratio within the landll body caused by the heterogeneity of waste, the heterogeneous character of the landll produced by compaction processes and resulting preferred water paths within the landll body are neglected. Doberl et al. (2003) suggested that leachate concentration is merely a result of the areas surrounding the preferential ow paths. The degree of heterogeneity together with the leachate amount determines the discharged substance load. The most detailed remarks on the approaches and backgrounds for calculations of multi-input inventory tools are given in Doka, 2003. The Ecoinvent database refers, as far as possible, to the specic chemical composition of the waste material (waste-specic burdens). In general in multi-inputmulti-output inventory tools, leachate emissions from specic elements in landlls (Ee) can be calculated as a function of the mass of the element e in waste fraction (me), the waste specic decomposition or degradability rate (D) and the element specic release factor (re), as there is discrepancy between the amount of waste that is decomposed and actual emission from the landll (e.g., Doka, 2003; Bjarnadottir et al., 2002). E e m D r e e 1

G. Obersteiner et al. / Waste Management 27 (2007) S58S74

S63

Additionally, the amount of gaseous emissions from the specic element must be subtracted. Data concerning the waste specic decomposition rate (D) can be derived from the literature (e.g., Micales and Skog, 1997; Zimmermann et al., 1996; Lechner, 2004). The lack of exact knowledge of the transfer mechanisms during waste treatment makes it necessary to apply dierent assumptions to allocate emissions to inputs. The crucial point in modelling is the element specic release factor. Doka, 2003 in the most recent Life Cycle Inventory for landlls resorts to empirical data by calculating the factor based on the mean annual leachate output from generic landll data. Of course, Doka is aware of the fact that landll sites are very heterogeneous regarding the composition of waste and development of the landll depending on various conditions (e.g., climate). Although this coarse estimate in general should provide tolerable results, not only the developer but also all users of similar multi-inputmulti-output inventory tools should be aware that several elements may be misjudged and that moreover, all assumptions are based on specic empirical data that may vary as to general conditions (rainfall, temperature, age of the landll when measurements were conducted, height of the landll, waste composition etc.). Currently, only chemical elements such as copper, zinc, and nitrogen are taken into account by disposal models as Ecoinvent 2000. Chemical compounds such as dioxins or other hydrocarbons have not been included to date. The question is whether the methodologies actually applied for modelling are generally applicable to the whole periodic table of elements of chemical compounds at all. An alternative method to model the complex heterogeneous environment of a landll is introduced by Zacharof and Butler (2004). As the microbial processes govern the breakdown of the waste and the hydrological processes control the water and leachate movement, they couple a simplied microbial degradation model with a stochastic hydrological and contaminant transport model. In view of the data gap, particularly with concern for longterm emissions from landlls, one has to keep in mind that the goal precision of leaching test studies to predict longterm emissions is higher than one magnitude. The Institute of Waste Management at BOKU-University developed a combined multi-phase weathering and leaching procedure (MPWLP test) for the expanded basic characterisation of the leaching behaviour of inorganic waste (Sabbas, 2001). Under the MPWLP-procedure, mineralogical changes that signicantly inuence heavy metal emissions, especially carbonation, are accelerated in an articial laboratory weathering step prior to leaching (Mostbauer et al., 2003). All multi-input inventory tools present the common factor of presuming the disposal of specic single waste fractions (not just average waste). Among the many areas of waste management, waste disposal represents only a part of the entire waste management system. The exact composition of the waste is not known. Local waste management systems and local resource management systems generally display specic characteristics such as composition and ori-

gin of raw materials, emissions from landlls or incineration plants, or energy sources used for electricity supply. Specic data on waste composition are dependent on local waste management systems. For example separate collection of plastic may include all plastic packaging including composites or may include plastic bottles only. Depending on the local system the separate collection of a specic waste fraction therefore does not allow conclusions to be drawn as to the composition of residual waste or collected fraction. In order to obtain specic data on waste composition, sorting analyses should be carried out. Depending on the composition and origin of raw materials, the chemical composition of waste materials may also vary greatly. However, in the increasingly globalised economies in which products and waste are imported and exported, the collection of such specic information is dicult and time-consuming (http://waste.eionet.eu.int/lca). In the highly developed western European or northern American countries, the drawing of conclusions from other countries may be acceptable because of similar threshold values as well as roughly similar external factors such as climate conditions. With regard to developing countries or new EU member states, the characterisation of waste as a mixture of various materials and therefore the generation of waste specic data may be dicult. One should be aware that the respective mobilisation (re-precipitation) rate of individual elements depends on the amount of newly accumulated leachate, as well as on the bulk density or the height of the landll. Calculations and prognosis can only be made for a specic leachate generation rate and therefore a specic climate and a specic height of the landll. Existing tools for waste management including landlls in general do not always allow for consideration of specic landll circumstances such as climate, average rainfall, height and type of cover layer, density, and permeability. Especially in the case of landlls where the specic process conditions are often unknown and cannot be controlled, it seems to be impossible to develop predictive models. This is due to the fact that landlling is not a single process but rather a series of independent processes (Van der Ven, 1997). Under these circumstances, the model cannot provide an extensive reproduction of reality. The complexity of landlls necessitates a multitude of assumptions and restrictions in modelling. This may be one reason why at present no multi-input inventory tools for MBT landlls have been developed. It may be concluded that neither general methodology (empirical results and modelling) can be preferred as the only right solution. One must be aware that both methodologies may underestimate or overestimate the actual or future concentrations of leachate in a landll. Therefore the waste composition and associated maximum possible pollution should be borne to mind. As already determined for the choice of time frame, before opting for one specic methodology the goal of the study and the practicability of the methodology should be taken into account. Dierent

S64

G. Obersteiner et al. / Waste Management 27 (2007) S58S74

decisions may be taken for the comparison of dierent waste management strategies or the life-cycle assessment of one specic product. 3. System characterisation As discussed in the previous chapters, the providing of a good database for landlls is a very dicult task. Therefore, in the following database several restrictions should be pointed out. The main socio-environmental impacts arising from landlling can be summarised as followed: 1. the emission of greenhouse, ammable and noxious gases, 2. the leaching of nutrients, heavy metals and other toxic compounds with the potential to pollute surface and groundwater, 3. the generation of noise, dust, odours and other vectors, 4. change of the natural scenery, 5. the potential to cause deleterious health eects within local communities, 6. an increase in road trac. The discussion in this paper only aects the rst two points of this list, as these are the only points to cause methodological problems dealing with multi-input and multi-output processes and emissions occurring over a very long period of time. When taking into consideration developments over the last 100 years, predictions for the next millennia seem to be particularly uncertain. Therefore, although the authors are aware of the fact that within this time period the remaining waste will generally not be inert, in the following only a time horizon of 100 years will be taken into consideration. 3.1. Characterisation of analysed landll types A further restriction has been made with regard to the type of landlls addressed in this paper. Only non-hazard-

ous waste landlls are taken into account. Table 1 provides an overview of landll classication in the European Union, Austria, Germany and Switzerland and their allocation to each other. In addition to current landll types, former landlls that are no longer authorised according to the new EU Directive 1999/31/EC or country specic legislation but which still exist, are cited. For the eco-inventory the grey coloured landll types were included. Already in Austria and many EU member states and in other EU member states in the near future landlls may only accept waste that has either been pre-treated by incineration in order to attain a TOC of less than 5% or has undergone mechanical biological treatment. Obviously former landll types such as sanitary landlls and open dumps present as old deposits in Austria as well as in Europe, or which may even still be installed in some parts of Europe should be included in the following considerations. To include landlling in an LCA approach entails several methodological questions as shown in Section 2. Additionally, for the relatively new types of landll (MBT landll, landll for bottom ash from MSWI) no experiences are available for the mid-term behaviour (decades). Thus, for our analysis, a review of the literature was performed to evaluate both data from lab-scale tests as well as larger scale tests. Additional experience and data from our own investigations of MBT landlls and results from MPWLP tests (Binner, 2003; Mostbauer et al., 2003; Wurz, 1999) were included. Only references with proper meta information were used for the Life Cycle Inventory. 3.1.1. Landlling of untreated waste When municipal solid waste is landlled directly, anaerobic biological degradation produces landll gas and leachate. Over 90% of the converted organic carbon is released as CO2 and CH4. The remainder is released in the leachate (Binner, 2003).

Table 1 Former and actual Landll types in Europe, Austria, Germany and Switzerland EU Open dumps Sanitary landlls Landlls for hazardous waste Landlls for nonhazardous waste Austria Open dumps Sanitary landlls Massenabfalldeponie Reststo-deponie Bodenaushubdeponie Baurestmassendeponie Germany Open dumps Sanitary landlls Landll class III Landll class II Landll class I Landll class 0 Residual material landlls Inert material landlls Switzerland Open dumps Sanitary landlls Explanation Landlls for untreated municipal waste, without specic technical safety measures Landlls for untreated municipal waste, with physical barriers to protect the public from waste Landlls for hazardous waste Landlls for MBT waste Landll for MSWI residues Landll for excavated earth Landll for construction and demolition waste

Landlls for inert waste

G. Obersteiner et al. / Waste Management 27 (2007) S58S74

S65

3.1.2. Impact of the mechanical biological pre-treatment on landll behaviour In Austria it was possible to gain valuable knowledge about the long-term behaviour of wastes that had been mechanically and biologically treated (Binner, 2003). Impacts on the landll volume as well as on the water budget, the leachate composition and the landll gas emission are known. The treated waste had a placement density of 1.3 ton/m3, saving 3050% of landll volume. If sieve residues are incinerated, calculations showed that it is possible to save up to 77% of landll space (Raninger, 1995). The water budget of a landll is strongly inuenced by local conditions. However, investigations showed that because of a remarkably low permeability only a very low amount of leachate can be expected. Fehrer (2002) observed extremely low permeability in two dierent samples of pre-treated wastes (kf < 1010 m/s resp. kf < 1011 m/s). Leachate analyses proved that the acidication phase is eliminated by mechanical biological pre-treated waste (Binner, 2003). Leachate concentrations are in a lower range of concentrations than normally found in reactor type landlls. Binner (2003) concludes that MBT landlls show 9095% less organic burden than other landlls of the same age lled with untreated waste. Heavy metal concentrations were also reduced. The duration of pre-treatment has a very signicant impact on the generation of landll gas. Compared to conventional landlling of untreated waste, a reduction of gas production of up to 95% can be reached. 3.1.3. Impact of municipal waste incineration on landll behaviour After municipal waste incineration, the remaining residues shall not have more than 3%wt. of organic carbon (TOC). In contrast to landlls of untreated municipal waste among MSWI landlls, low landll gas formation can be expected because of the low organic content. Even though one cannot exclude gas generation entirely, because of the small database, no statements about amount and time dependent behaviour of gaseous emissions from MSWI landlls can be made (Kabbe, 2000). Decisive for the release and immobilisation of xed heavy metals in incineration residues is the development of pH value and redox potential of the landll body. Typically the pH-value from the leachate from MSWI slag in the rst decades after deposit is about 1011. Because of reactions between the alkaline components of the slag with carbon dioxide from the air (or from landll gas) and due to other weathering processes, the pH-value will decrease to values between 7.0 and 8.5. At the same time the LS ratio increases. In total a decrease (wash out) or stabilisation of the annual load (solubility control) of inorganic pollutants occurs. If the slag cannot react with carbon dioxide in a sucient way, e.g., because of a very fast buildup of the landll body, a strong mobilisation of aluminium, zinc and lead may occur. Altogether in MSWI landlls, a discharge of increased pollutant concentrations can be

expected at least occasionally (Hirschmann, 1999). Chemical treatment of leachate is necessary in most cases because of the initial alkalinity and heavy metal content. 3.2. Characterisation of environmental impacts of landlls 3.2.1. Landll leachate As for the leachate, the following parameters have to be borne in mind: The accumulating amount of leachate for all types of landll depends on the same factors that can be summed up in meteorology, latitude and altitude, vegetation, material properties and morphological factors. Normally the amount of leachate is given as a percentage of rainfall. It should be considered that the amount of rainfall inuences the resulting amount of leachate. If the amount of leachate is approx. 50% in regions with 1000 mm of annual rainfall, this is not the same for regions with 500 mm of annual rainfall. In regions with lower rainfall, the amount evaporating on the surface is comparatively higher than in regions with higher rainfall. As additional factors to rainfall and climate (temperature dependency of evaporation), the landll geometry constituted by height and density as well as surface and landll engineering (placement technology, cover, leachate management system) have to be considered (Binner, 2003). As an example, reports published for the amount of leachate for reactor landlls range from 25% to 60% of rainfall (Krumpelbeck, 2000; Plinke et al., 2000; Wallmann, 1999). According to Rettenberger and Fricke (1998), at optimised landll operation of MBT landlls the amount of leachate may fall below the area of 913% of rainfall. The quality of leachate depends on the toxic potential of the landlled material, the amount of leachate and the age and related age-dependant environmental factors of the landll. The formation of preferential ow paths was monitored for the disposal of solidied wastes for dierent landll types. The formation of cracks may occur, e.g., because of settlements and later on lead to the formation of water channels. The therewith associated changes in the water balance of landlls are dicult to predict. A consequence may be that water in the remaining, bypassed parts of the system must ow considerably slower, or not at all (Rosqvist and Destouni, 2000). One can assume as well that a dilution of leachate occurs after intensive rain due to this preferential ow. This fact has to be considered at the interpretation of leachate data. Due to this considerable variation in leachate, it is important that the system studied reects the actual technology applied in the geographical and temporal scope of the study. The description of medium and long-term behaviour of landlls by means of simple recipe like models seems to not be possible. Therefore assumptions have to be made. 3.2.2. Landll gas emissions The predicted amount of gas emissions from landlls can be estimated by the content of organic matter in waste.

S66 Table 2 Data on gas potential in the literature Source Stegmann and Dernbach (1982) Tabasaran and Rettenberger (1987) Plinke et al. (2000) Krumpelbeck (2000) DM, dry matter; FM, fresh matter.

G. Obersteiner et al. / Waste Management 27 (2007) S58S74

Gas potential 150200 m /ton 120300 m3/ton 150190 m3/ton 180280 m3/ton
3

Remarks DM (approx. 90120 m /ton FM) FM FM DM (approx. 100170 m3/ton FM)


3

Experimental Laboratory experiments Data from dierent landlls

The direct landlling of municipal solid waste is expected to have a gas potential from 100 to 180 m3/ton DM (dry matter) (Table 2). The gas potential related to the disposed waste (fresh matter, FM) is of importance in calculations. Table 3 gives an overview of data on gas potential found in the literature. According to investigations carried out by our group on solid municipal waste in the province of Salzburg, a gas potential of 120 m3/ton FM was used for the modelling in this paper (Binner et al., 1999). By providing a suitable mechanical processing of waste and optimised control of rotting process by aerobic pretreatment, a decrease of gas potential to about 510% basic value can be reached (Hei-Ziegler et al., 2004; Soyez, 2001). Measures on an MBT-landll in Austria showed a gas potential of 11.521.5 m3/ton (Rolland, 2001). In our study
Table 3 Specications concerning to used landll references Source Krumpelbeck (2000) Location The old West German states Type of material landlled High amount of household waste, as well as construction waste, contaminated soil and sewage sludge 80% Mechanical biological pretreated waste, 10% bulky, demolition and commercial waste; operation period 19801983; landll closed 1983 MBT-material n.s. MSWI bottom ash monoll, lled in 1992; not covered yet

according to the RA4 (respiration activity) limit volume of the Austrian landll directive (RA4 < 7 mg O2/g DM) and to our own investigations of pre-treated material, a gas potential of 15.6 m3/ton FM was calculated (Table 4). 4. Empirical LCI data for dierent landll types in Central Europe In the following inventory, data for the four dierent landll types under investigation are presented for the three dierent time periods. Data are from the literature and our own investigations. In view of the fact that in leachate composition, water represents the transport medium for pollutants it is important to focus on the sources, factors and processes which govern the discharge. Climatic conditions

Database 36.780 leachate analysis, 3.970 gas analysis

Comment Evaluation and assessment of 76 landlls; less than half of them with landll bottom liner, most landlls with leachate and gas collection system; average annual values

Wurz (1999)

AttnangRedlham, Austria, Wilhelmshaven, Germany Lostorf near Buchs, Switzerland Vestskoven, Denmark

Turk et al. (1997) Johnson et al. (1999) Hjelmar (1995)

Longstanding series of leachate analysis; data from 1980 to 1983 (Phase 1) and 1984 to 1992 (Phase 2) n.s. Average values from 194 samples during experimentation period from November 28, 1994 to November 15, 1996 Annual analysis of leachate 19731992, 22 observations; average values for 1973/1974 (phase 1) and 1991/1992 (phase 2) Landll age about 27 years DSR values

Values for AOX, Na and K for Period II of MBT landlls Values for average conditions excluding rain and dry periods

MSWI ash monoll, containing 10,000 tons of bottom ash with 15% y ash

Values for 02 year old landll and 1819 year old landll;

Hjelmar (1995) Kruse (1994) and Kabbe (2000) Zweifel et al. (1999) Mostbauer (2005)

Woodburn, Oregon US Laboratory values

Combined ash monoll Untreated residual waste

Only maximum values for Na, K, Ca from Woodburn n.s.

Switzerland, Riet Rautenweg, Vienna, Austria

Mainly household waste, landlled 19251935 Weathered incineration residues; MSWI bottom ash and MSWI bottom ash containing some other ashes

Longstanding series of leachate analysis; data from 1994 to 1997 Mean value from leachate analysis between 1999 and 2003 from Rautenweg test cells;

Compare Jaros and Huber (1997), Sabbas et al. (2001) and Mostbauer (2005) for details about the laboratory based work in the eld of weathering of MSWI bottom ash

n.s., not specied.

G. Obersteiner et al. / Waste Management 27 (2007) S58S74 Table 4 Summary of assumptions for landll models Open dump Annual landll rate (t) Average landll height Bulk density Bottom layer Cover layer Operation period (I) Active aftercare period (II) Modelling period (III) Landll gas potential Period I Period II Period III Active landll gas collection Period I Period II Period III Methane oxidation rate (Periods II and III) Average annual rainfall Amount of leachate [% of rainfall] Period I Period II Period III Leachate treatment Period I Period II Period III 242,958/104,483 10 m 1 ton/m3 No No 5 years 30 years 100 years 120 m3/ton FM 22% 75% 3% No 120 m3/ton FM 22% 75% 3% No 45% No 90% 1300 mm 50% 30% 30% 15.6 m3/ton FM 30% 60% 10% No
a

S67

Sanitary landll 104,483 1 ton/m3 State of the art Methane oxidation layer

MBT-landll 29,315 1.4 ton/m3 State of the art Methane oxidation layer

MSWI-landll 26,762 1.5 ton/m3 State of the art Recultivation layer

No 1300 mm 50% 50% 50% No

90% 1300 mm 50% 5%/30%a 5%/30%a

1300 mm 65% 30% 30% PF-AC-Ox PF-AC-Ox

B-MF-RO RO B-MF-RO RO No (impermeability of bottom layer uncertain)

B-MF-RO: Biological treatment, MicroFiltration, Reverse Osmosis. PF-AC-Ox: Precipitation/Flocculation, Activated Carbon lter, Oxidation. a Two variations considered.

(precipitation characteristics, temperature and wind), vegetation and the type of surface soil determine the water balance. The discharge pattern also depends on the permeability of the material, as well as the possible presence of preferential path ows. Johnson et al. (1998) proved that in the presence of preferential ow paths, the discharge from a MSWI bottom ash landll is characterised by long periods of low and nearly constant ows interspersed with increases in leachate quantity after rain. With a view to landll gas composition, the cover layer and type of surface, respectively, are particularly important as determining factors. Engineered methane oxidation covers are very eective in reducing greenhouse gas emissions from landlls. Water retention layers and well-planned recultivation at the surface may reduce leachate amounts signicantly. Therefore, when publishing landll emission data or working with landll data from the literature, it is important to know as much as possible about the metadata, the background conditions responsible for the specic landll emissions. Beside the general climate conditions and the age of the landll, it is fundamental to ascertain as far as possible the generation of data pertaining to leachate (average data over a certain time period, quantity of measurements, single analysis during rainy or dry periods, etc.).

Table 3 provides a review of the references used in preparation of the following inventory tables. In the present paper it is assumed that, with the exception of Open Dumps, a fully functional leachate treatment has been implemented during the rst 30 years following landll construction, and that 100% of the leachate is collected and treated in this active aftercare period. In the opinion of the authors, the impermeability of the bottom liner system can no longer be guaranteed after this period, after which direct leachate emissions into groundwater are to be expected. To consider the leachate emissions into groundwater, the eventual presence of geological barriers has not been taken into consideration and therefore has to be taken into account separately. To take into account dierent quantities and qualities of leachate in view of the age of the landll, emissions were considered separately for the periods 05 years (operation period), 630 years (active aftercare period) and 31100 years (medium time period, where active aftercare has stopped). All assumptions for modelling are summarised in Table 4. After a detailed literature survey, values were taken only from studies where detailed meta-information concerning the landlls and generated data was given (Table 3). Tables 510 provide detailed information on the life cycle inventory of dierent types of landlls. They are

S68

G. Obersteiner et al. / Waste Management 27 (2007) S58S74

Table 5 Leachate concentration for dierent landll types during operational period I (year 15); in cases when values reported in the literature exceeded legal limit values and legal limit values were used, these values are marked in grey

based on a black box concept, meaning that depending on the amount of waste landlled, the specic emissions can be calculated from the concentration of the specic element in the leachate or gas and the amount of leachate or gas developed. In view of the fact that for sanitary landll, MBT landll and MSWI landll leachate, treatment is obligatory throughout the rst 30 years, the specic emission values have been replaced by threshold values according to the Austrian sewage emission ordinance (Abwasseremissionsverordnung, BGBl 263/2003; BGBl 179/1991; BGBl.Nr. 186/1996) in cases when values reported in the literature exceeded this legal limit value. These values are marked in grey (Tables 57). Additionally, expenditures incurred in leachate treatment (e.g., electricity) should be calculated, being identical for both landll operation or for the use of energy from collected landll gas. These gures should be calculated according to the specic conditions under investigation. Tables 57 show the concentration of pollutants in the leachate of the four dierent landll types investigated. In order to avoid any attempt at obtaining a non-existent

degree of accuracy, values are rounded. Empty elds indicate that no useable values were found. Zero indicates that measured values are below limits of detection. To calculate the leachate emission potential for the rst 100 years, the leachate ow rate according to Eq. (2) should be applied X Eleach; 100 Rf Leach%I cLeach; I Rf Leach%II cLeach; II Rf Leach%III cLeach III 2 Eleach, 100 is the leachate emission for element e in the rst 100 years of landll, Rf is average annual rainfall (ml), Leach%I is amount of leachate in percent of the rainfall in periods I, II or III from Table 4 (%), cLeach I to III is leachate concentration of the element e for Periods I, II or III from Tables 5,6 or 7. Tables 810 show the concentration of pollutants in the leachate of the four dierent landll types investigated. The process of calculation is slightly dierent, starting from the overall landll gas potential. To calculate the gas emission potential for periods II and III (year 6100),

G. Obersteiner et al. / Waste Management 27 (2007) S58S74

S69

Table 6 Leachate concentration for dierent landll types during active aftercare period II (year 630); in cases when values reported in the literature exceeded legal limit values and legal limit values were used, these values are marked in grey

active gas collection and methane oxidation should also be taken into consideration, as illustrated in Tables 9 and 10. 5. Summary and conclusions Not only the production and the use of materials can produce a burden to the environment but also the treatment and disposal of the same products as waste provides a far from negligible contribution to environmental pollution. It is therefore common sense that the disposal phase should by default be included in LCA studies. Both from a waste management and a life-cycle assessment point of view it is clear that disposal, and therefore the emissions occurring from landlls, should be included in the evaluation of waste management options as well as in the assessment of a products life cycle. As discussed in the present paper the above consideration leads to several methodological questions. The main questions debated are related to the issues of time frame and data generation. It is common sense that the use of long time frames (innite, if possible) and the use of

calculation methodologies taking into account the specic waste composition should be applied. Both aspects lead to time-consuming and therefore costly procedures. In contrast to other processes involved in the production or management of waste where the emissions from the process can be measured directly and immediately, this is not true for landlls. Landll emissions do not occur at one point (like a smokestack) but there are diuse emissions manifested over a very long time span. A direct measurement of pollutant loads occurring over many decades is not possible. Therefore assumptions have to be made. When including landll emissions in an assessment procedure one is confronted with numerous uncertainties on the one hand and very often with a complicated calculation procedure on the other. This paper attempts to demonstrate how the latter depends particularly on the goal and content of the specic LCA, and to question whether it is really necessary or even useful to work with long time horizons exceeding the calculable time period of 100 years and/or to calculate waste specic emissions by means of multi-inputmulti-output tools. It is demonstrated that although for many issues such as comparison of recycling,

S70

G. Obersteiner et al. / Waste Management 27 (2007) S58S74

Table 7 Leachate concentration for dierent landll types during foreseeable time period III (year 31100) Landll type Open dump Mean pH-value TOC BOD5 COD NH4N NO2N NO3N SO4 SO3 Cd Fe Zn AOX Na K Ca Mg Mn Pb Cu Ni Cr Hg As Al Sb B Ba Co Mo Si V mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l 7.70 500 300 1200 120 0.84 10.00 80 1.1 0.0028 12.50 0.54 1.13 600 600 200 100 0.91 0.03 0.04 0.12 0.18 0.04 Sanitary landll Mean 7.70 500 300 1200 120 0.84 10.00 80 1.1 0.0028 12.50 0.54 1.13 600 600 200 100 0.91 0.03 0.04 0.12 0.18 0.04 MBP landll Mean 7.70 500 300 1200 120 0.84 10 80 1 0.0028 13 0.54 1 600 600 200 100 1 0.03 0.04 0.12 0.18 0.04 MSWI landll Mean 7.60

2.900 0.0027 0.09 0.13 500 200 600 400 0.07 0.0027 0.03 0.03 0.01 0.0004 0.0047 0.05 0.03 3.27 0.03 0.0017 0.26 4 0.0043

10.00

10.00

10

Table 8 Landll gas concentration for dierent landll types during operational period I (year 15) Landll type Landll gas potential Proportion year 15 Landll gas potential year 15 CH4 CO2 CH4 CO2 m3/ton % m3/ton Vol% Vol% m3/ton m3/ton Min CH4 CO2 Benzene Toluene o-Xylol p/m Xylol Dichloromethane Trichloromethane Tetrachloromethane 1,1,1-Trichloroethane Trichloroethene Tetrachloroethene g/t g/t mg/ton mg/ton mg/ton mg/ton mg/ton mg/ton mg/ton mg/ton mg/ton mg/ton 11,339 20,738 0.26 0.53 5.28 0 0 0 0 0.013 0 0.003 Open dump 120 22% 26.4 60 40 15.8 10.6 Max 11,339 20,738 2508 16,236 185 9926 6600 52.8 15.8 726.0 4804.8 3748.8 Min 11,339 20,738 0.3 0.5 5.3 0 0 0 0 0.013 0 0.003 Sanitary landll 120 22% 26.4 60 40 15.8 10.6 Max 11,339 20,738 2508 16,236 185 9926 6600 52.8 15.8 726.0 4804.8 3748.8 Min 2010 3676 0.9 0.1 0.1 0.3 0.017 0.002 0.002 0.003 0.014 0.005 MBP landll 15.6 30% 4.68 60 40 2.8 1.9 Max 2010 3676 3.7 13.5 12.6 32.5 0.327 0.006 0.002 0.042 0.050 0.032

G. Obersteiner et al. / Waste Management 27 (2007) S58S74 Table 9 Landll gas concentration for dierent landll types during active aftercare period II (year 630) Landll type Landll gas potential Proportion year 630 Landll gas potential year 630 Active gas collection Gas used for energy production Landll gas emissions year 630 CH4 CO2 CH4 CO2 Methane oxidation CH4-emissions CO2-emissions m /ton % m3/ton % m3/ton m3/ton Vol% Vol% m3/ton m3/ton % % m3/ton m3/ton Min CH4 CO2 Benzene Toluene o-Xylol p/m Xylol Dichloromethane Trichloromethane Tetrachloromethane 1,1,1-Trichloroethane Trichloroethene Tetrachloroethene g/ton g/ton mg/ton mg/ton mg/ton mg/ton mg/ton mg/ton mg/ton mg/ton mg/ton mg/ton 38,657 70,698 0.9 1.8 18 Max 38,657 70,698 8550 55,350 630 33,840 22,500 180 54 2475 16380 12,780
3

S71

Open dump 120 75% 90

Sanitary landll 120 75% 90 45% 40.5 49.5 60 40 29.7 90 19.8 2.97 46.53 Min 2126 91,378 0.495 0.99 9.9 Max 2126 91,378 4703 30,443 347 18,612 12,375 99 30 1361 9009 7029 Min 402 17,279 1.760 0.140 0.197 0.599 0.034 0.004 0.003 0.007 0.028 0.010

MBP landll 15.6 60% 9.36

90 60 40 54 36 54 36

9.36 60 40 5.62 90 3.74 0.562 8.798 Max 402 17,279 7.413 26.947 25.132 64.977 0.654 0.012 0.003 0.083 0.099 0.065

0.045 0.009

0.025 0.005

Table 10 Landll gas concentration for dierent landll types during foreseeable time period III (year 31100) Landll type Landll gas potential Proportion year 30100 Landll gas potential year 630 CH4 CO2 CH4 CO2 Methane Oxidation CH4-emissions CO2-emissions m /ton % m3/ton Vol% Vol% m3/ton m3/ton % % m3/ton m3/ton Min CH4 CO2 Benzene Toluene o-Xylol p/m Xylol Dichloromethane Trichloromethane Tetrachloromethane 1,1,1-Trichloroethane Trichloroethene Tetrachloroethene g/ton g/ton mg/ton mg/ton mg/ton mg/ton mg/ton mg/ton mg/ton mg/ton mg/ton mg/ton 1546 2828 0.036 0.072 0.72 0 0 0 0 0.002 0 0.0004
3

Open dump 120 3% 3.6 60 40 2.16 0 1.44 0 2.16 1.44 Max 1546 2828 342 2214 25.2 1354 900 7.2 2.16 99 655 511

Sanitary landll 120 3% 3.6 60 40 2.16 90 1.44 90 0.216 3.384 Min 155 6646 0.036 0.072 0.72 0 0 0 0 0.002 0 0.0004 Max

MBP landll 15.6 10% 1.56 60 40 0.936 90 0.624 90 0.094 1.466 Min 67 2880 0.293 0.023 0.033 0.100 0.006 0.001 0.001 0.001 0.005 0.002 Max 67 2880 1.236 4.491 4.189 10.830 0.109 0.002 0.001 0.014 0.017 0.011

155 6646 342 2214 25.2 1354 900 7.2 2.16 99 655 511

incineration and landlling as alternatives the choice of the time horizon may represent a most important factor, with regard to other issues such as treatment of municipal solid waste in MBT, incineration or direct landlling as alternatives such long-term emissions no longer have any eect.

This is due mainly to the reason that for dierent landll types the amount of pollutants released into water and air diers according to time. These dierences however only play a signicant role in the rst 100 years. The same applies for the choice of methodology, where the decision

S72

G. Obersteiner et al. / Waste Management 27 (2007) S58S74

should be made on the basis of the goal of the study as well as on the resulting system boundaries. When carrying out a product LCA we usually wish to establish some kind of causal relations between a specic product or material and the emissions actually caused by the product studied (Sundqvist et al., 1997). If a waste management system (starting from waste generation not from production) is the object of investigation, even raw data of waste composition may be lacking and indeed at times be superuous in cases where the same type of waste is treated and landlled in dierent ways. The option of considering the most extreme case (all substances contained in deposited waste will be released from the landll sooner or later) in which the chemical composition of waste is known, constitutes an easy to use alternative. Further to performing a detailed examination of methodological issues the second goal of this paper was to provide a useful database of landll emissions for the central European region. The inventory compiled from emission data from four dierent landll types for three dierent periods of time provides a detailed data base for direct use wherever general conditions apply to the conditions described in the paper, as well as for further calculations for universally valid modelling e.g. for MBT landlls. A considerable challenge to the use of LCA in waste and resource management is represented by the fact that the impact of these systems depends largely on regional conditions, including consumer habits, mode of transport, generation of by-products and energy, or the energy supply systems in place (fossil fuels, biomass, hydropower, nuclear, wind). It is therefore important that when possible local data be applied, rather than importing external data or using the default data. As a conclusion one can identify three dierent areas where the focus of further investigations should be laid on. As outlined in the paper one big problem for the assessment of the impact of landlls within LCA is still the lack of data. Both specic emission data as well as corresponding meta data have to be investigated in long-term studies. Special interest should be given to chemical compounds such as dioxins which often are not included into existing databases. This is mainly due to the fact that the content of those compounds in leachate is often near the limit of detection and therefore a calibration of models is dicult or even impossible. But not to include those substances in the LCA may aect the result because of their e.g. high human toxicity potential. In general in the last decades a lot of models have been developed. For the future it seems necessary that such models have to be subject to a sort of quality assurance which includes: (a) verication of the assumptions and the program code, (b) calibration with eld data or data from large pilot plants and (c) external evaluation. As a third area of interest one has to detect the relevance of consideration of landlls in LCA-analysis of disposal systems for dierent scenarios to give a guideline for fur-

ther investigations under which conditions landlls have to be included and in which degree of detail.

References
Binner, E., 2003. MBT with view to landlling stabilised biowaste. In: H. Langenkamp (Soil and Waste Unit, Institute for Environment and Sustainability, Joint Research Centre), Luca Marmo (Sustainable Resources Unit, Directorate-General Environment, European Commission): Workshop Biological Treatment of Biogradable Waste Technical Aspects, 810 April, 2002, Brussels: 2003, EUR 20517 EN, p. 355 f. Binner, E., 2003. The impact of mechanicalbiological pretreatment on the landll behaviour of solid wastes. In: Langenkamp, H., Marmo, L. (Eds.), Biological Treatment of Biodegradable Waste Technical Aspects. Workshop Proceedings, Brussels. Binner, E., Zach, A., Lechner, P., 1999. Stabilitatskriterien zur Charak terisierung der Endprodukte aus MBA - Anlagen. Studie im Auftrag des Bundesministerium fur Umwelt, Jugend und Familie, Wien. Bjarnadottir, H., Fririksson, G., Johnsen, T., Sletsen, H., 2002. Guidelines for the use of LCA in the waste management sector, Nordtest Report TR 517. Coleman, T., Masoni, P., Dryer, A., McDougall, F., 2003. International Expert Group on Life Cycle Assessment for Integrated Waste Management. International Journal of LCA 8 (3), 175178. Doberl, G., Huber, R., Fellner, J., Brunner P.H., 2003. The heterogeneity of waste as a main factor inuencing the future emission potential of MSW landlls a case study on the Breitenau landll in Austria. In: Christensen, T.H., Cossu, R., Stegmann, R. (Eds.), SARDINIA 2003 Ninth International Waste Management and Landll Symposium, 6 10 October 2003, S. Margherita di Pula (Cagliari), Italy; Proceedings (Abstract; long version on CD). Doka, G., 2003. Life Cycle Inventories of Waste Treatment Services. Part III: Landlls Underground Deposits Landfarming, BUWAL. Doka, G., 2003. Landlls underground deposits landfarming, part III. Life Cycle Inventories of Waste Treatment Services. Ecoinvent Report No. 13, Swiss Centre for Life Cycle Inventories, Dubendorf, December 2003. Doka, G., 2004. Modelling of long-term emissions in LCIs of landlls. Presentation at the 22. LCA discussion forum Evaluation of Long-Term Impacts in LCA, May 7th, 2004, ETH Zurich, Switzerland. Available from: <http://texma.org/lcaforum/Documentation/DokaDF22PUBLIC.pdf>. Doka, G., 2005. Assessing long-term eects of municipal solid waste. In: Lechner, P. (Ed.), Waste Management in the Focus of Controversial Interests. Proceedings of 1st BOKU Waste Conference. Facultas Verlags- und Buchhandels AG. Doka, G., Hischier, R., 2005. Waste treatment and assessment of longterm emissions. Int J LCA 10 (1), 7784. Eriksson, O., Frostell, B., Bjorklund, A., Assefa, G., Sundqvist, J.-O., Granath, J., Carlsson, M., Baky, A., Thyselius, L., 2002. ORWARE a simulation tool for waste management. Resources, Conservation and Recycling 36, 287307. EU Directive 1999/31/EC: Council Directive 1999/31/EC of 26. April 1999. Ocial Journal of the European Community, L 182/1. Fehrer, K., 2002. Geotechnisches Verhalten von MBA-Material, Diplomarbeit am Institut fur Abfallwirtschaft der Universitat fur Bode nkultur, Wien. Finnveden G., 1998. On the possibilities of life-cycle assessment development of methodology and review of case studies. Doctoral Thesis in Natural Resources Management, Department of Systems Ecology, Stockholm University. Finnveden, G., 1999. Methodological aspects of life cycle assessment of integrated solid waste management systems. Resources, Conservation and Recycling 26, 173187.

G. Obersteiner et al. / Waste Management 27 (2007) S58S74 Frischknecht, R., Rebitzer, G., 2005. The ecoinvent database system: a comprehensive web-based LCA database. Journal of Cleaner Production 13 (2005), 13371343. Frischknecht, R., Jungbluth, N., Althaus, H.-J., Doka, G., Dones, R., Hellweg, S., Hischier, R., Nemecek, T., Rebitzer, G., Spielmann, M., 2004. Code of practice, Ecoinvent Report No. 2, Data v1.1, Duben dorf. Hei-Ziegler C., Lechner P., Mostbauer P., 2004. Endlager Deponie. In: Lechner P. (Ed.), Kommunale Abfallentsorgung. UTB 2114, Facultas, Wien, ISBN 3-8252-2114-8, pp. 29135. Hellweg, S., 2000. Time- and Site-Dependent Life-Cycle Assessment of Thermal Waste Treatment Processes. Dissertation No. 13999, ETH, Zurich. Hellweg, S., Hofstetter, T., Hungerbuhler, K., 2003. Discounting and the environment. Should current impacts be weighted dierent than impacts harming future generations? International Journal of LCA 8 (1), 818. Hirschmann, G., 1999. Langzeitverhalten von Schlacken aus der thermischen Behandlung von Siedlungsabfallen.Fortschr.-Ber. VDI Reihe 15 Nr 220, Dusseldorf, VDI Verlag. Hjelmar, O., 1995. Incineration of MSW: implications for landll strategy and leachate management. In: Proceedings Sardinia 95, 5th International Landll Symposium. Jaros, M., Huber, H., 1997. Emissionsverhalten von MVA Schlacke nach forcierter Alterung. Waste Reports No. 06, Eigenverlag BOKU. Johnson, A., Richner, G.A., Vitvar, T., Schittli, N., Eberhard, M., 1998. Hydrological and geochemical factors aecting leachate composition in municipal solid waste incinerator bottom ash part 1: The hydrology of Lostorf, Switzerland. Journal of Contaminant Hydrology 33, 361376. Johnson, A., Kaeppeli, M., Brandenberger, S., Ulreich, A., Baumann, W., 1999. Hydrological and geochemical factors aecting leachate composition in municipal solid waste incinerator bottom ash. Part II: The geochemistry of leachate from Landll Lostorf, Switzerland. Journal of Contaminant Hydrology 40, 239259. Kabbe, G., 2000. Emissionsverhalten behandelter und unbehandelter Siedlungsabfalle wahrend der Ablagerung. Abfall-Recycling-Altlasten 21; Aachen 2000. Koblmuller, M., Konrad, W., Pladerer, C., Karner, P., Loidl, M., Salhofer, S., Wassermann, G., Binner, E., Stubenvoll, J., 2004. Strategische Umweltprufung Abfallwirtschaftsplan Salzburg, vorlau ger Umweltbericht, Salzburger Landesregierung, Abt. 16, Salzburg 2004. Krumpelbeck, I., 2000. Untersuchungen zum langfristigen Verhalten von Siedlungsabfalldeponien; Vero. des Lehrstuhls fur Abfall- und Siedlungswasserwirtschaft der Bergischen Universitat, Ges amthochschule Wuppertal, Wuppertal. Kruse K., 1994. Langfristiges Emissionsgeschehen von Siedlungsabfalldeponien. Dissertation, Fachbereich fur Bauingenieur- und Vermessung swesen, Technical University Braunschweig, Germany. Lechner P., 2001. Deponietechnik und Altlastensanierung, Vorlesungsunterlagen, Universitat fur Bodenkultur Wien; Abteilung Abfallwirts chaft, Austria. Lechner, P., 2004. Kommunale Abfallentsorgung. Facultas UTB. McDougall, F., White, P., Franke, M., Hindle, P., 2001. Integrated Solid Waste Management: A Life Cycle Inventory. Blackwell Science, London. Melloni, R., Carlini, N., Neri, P., Pozzi, V., Bergnzoni, M., 2003. Environmental and economic analysis of mechanical biological pretreatment of MSW: an LCA approach. In: Christensen, T., Cossu R., Stegmann R. (Hrsg.): Sardinia 2003 Ninth International Waste Management and Landll Symposium, 610 Oktober 2003, S. Margherita di Pula, Cagliari, Sardinien, Italien. Proceedings; CISA Environmental Sanitary Engineering Centre; pp. 615616 (Abstract; full paper on CD). Micales, J.A., Skog, K.E., 1997. The decomposition of forest products in landlls. International Biodeterioration and Biodegradation 39 (23), 145158.

S73

Mostbauer, P., 2005. Leachate composition. In: Lechner P. (Ed.), Waste Management in the Focus of Controversial Interests. 1st BOKU Waste Conference 2005. Mostbauer P., Riegler H., Lechner P., 2003. Articial weathering of inorganic waste evaluation of laboratory weathering and leaching methods. In: Christensen, T.H., Cossu, R., Stegmann, R. (Eds.), Sardinia 2003 Ninth International Waste Management and Landll Symposium, 610 Oktober 2003, S. Margherita di Pula (Cagliari), Italy; Proceedings (Abstract; Langfassung auf CD), pp. 288289. Nielsen, Per H., Hauschild, M., 1998. Product specic emissions from municipal solid waste landlls. Part I: Landll modell. International Journal of LCA (3), 158168. Plinke, E., Schonert, M., Meckel, H., Detzel, A., Giegrich, J., Fehrenbach, H., Ostermayer, A., Schorb, A., Heinisch, J., Luxenhofer, K., Schmitz, S., 2000. Okobilanz fur Getrankeverpackungen II. Umweltforschungs plan des Bundesministeriums fur Umwelt, Naturschutz und Reak torsicherheit im Auftrag des Umweltbundesamtes. Raninger, B., 1995. Sammlung und Verwertung kommunaler biogener Abfalle in Osterreich. Eigenverlag Bundesministerium fur Umwelt. Rettenberger, G., Fricke, K., 1998. Anforderung an die Deponierung von Restabfallen aus der mechanisch-biologischen Abfallbehandlung. In: Wiemer, K., Kern, M., (Eds.), Bio- und Restabfallbehandlung II, M.I.C. Baeza-Verlag, Witzenhausen. Rolland, C., 2001. Ablagerungsverhalten mechanisch-biologisch vorbehandelter Abfalle. Eigenverlag BMFLUW. Rosqvist, H., Destouni, G., 2000. Solute transport through preferential pathways in municipal solid waste. Journal of Contaminant Hydrology 46, 3960. Sabbas, T., 2001. Evaluierung von Testmethoden und Modellen fur die Bewertung des mittel- bis langfristigen Emissionsverhaltens schadstohaltiger anorganischer Abfalle und Immobili sierungsprodukte, im Auftrag der Stadt Wien, BMLFUW, BMBWK und INTERREG. Sabbas, T., Lechner, P., 2001. Developing an assessment method for the middle- to long-term pollutant ux of inorganic residues. In: Proceedings from the Solid Waste Association of North Americas 6th Annual Landll Symposium. Hyatt Regency Islandia on Mission Bay. San Diego, California, June 1820, 2001. Sabbas, T., Mostbauer, P., Lechner, P., 1998. Deponien Prozesse und Faktoren jenseits der Nachsorge, Proejtkbericht im Auftrag der MA 22 Umweltschutz. Sabbas, T., Mostbauer, P., Lechner, P., 1999. Final storage: time scale, processes and factors beyond aftercare. In: Cossu, R., Stegmann, R. (Eds.), Sardinia 1999. Proceedings of the 7th International Waste Management and Landll Symposium, Sardinia, vol. I, pp. 310. Sabbas, T., Mostbauer, P., Rautner, T., Lechner, P., 2001. A prediction model for the assessment of the middle and long term emission ux of inorganic Wastes. In: ISWA (Ed.), Proceedings of the ISWA World Congress 2001 Waste in a Competitive World, ISWA International Solid Waste Association, Denmark, pp. 8293. Soyez, K., 2001. Mechanisch biologische Abfallbehandlung. Erich Schmidt Verlag.. Stegmann R., Dernbach H., 1982. Deponieentgasung mittels Gasbrunnen zum Zwecke optimaler Gasnutzung am Beispiel der Deponie Braunschweig. Beiheft 19 zu Mull und Abfall, S. 3946. Sundqvist, J.-O., 1999. Life cycle assessment and solid waste guidelines for solid waste treatment and disposal in LCA. AFR Report 279. Sundqvist, J.-O., 2002. Some methodological questions and issues that are of great interest for the result in LCA. In: Sundqvist, J.-O., Finnveden, G., Sundberg, J. (Eds.), Proceedings from Workshop on System Studies of Integrated Solid Waste Management in Stockholm, 23 April 2001. IVL Report B1490. Sundqvist, J.-O., Finnveden, G., Albertsson, A.-C., Karlsson, S., Berendson, J., Hoglund, L.O., 1997. Life cycle assessment and solid waste stage 2. AFR-Report 173; AFR, Stockholm, Sweden. Tabasaran O., Rettenberger, G., 1987. Grundlagen zur Planung von Entgasungsanlagen, Hosel/Schneider/Schnurer Mullhandbuch Nr. 4547.

S74

G. Obersteiner et al. / Waste Management 27 (2007) S58S74 gie- und Schadgasbilanzen. Schriftenreihe d. Arbeitskreises fur die Nutzbarmachung von Siedlungsabfallen. Wassermann, G., Binner, E., Salhofer, S., 2005. Relevance of Consideration of landlls in LCA analysis of disposal. In: Lechner P. (Ed.), Waste Management in the Focus of Controversial Interests. 1st BOKU Waste Conference 2005. Wurz, H., 1999. Auswirkungen mechanisch-biologischer Restabfallbehandlung auf das Deponieverhalten. Diplomarbeit am Institut fur Abfallwirtschaft der Universitat fur Bodenkultur, Wien. Zacharof, A.I., Butler, A.P., 2004. Stochastic modelling of landll processes incorporating waste heterogeneity and data uncertainty. Waste Management Special Issue on Landll modelling 24 (3), 241250. Zimmermann, P., Doka, G., Huber, F., Labhardt, A., Menard, M., 1996. Okoinventare von Entsorgungsprozessen. Grundlagen zur Integration der Entsorgung in Okobilanzen. Eigenverlag ETH Zurich. Zweifel, H.-R., Johnson, C.A., Hoehn, E., 1999. Langzeitanalysen der Hauptelemente von Sickerwassern aus Altdeponien. Mull und Abfall (1999) Heft 12 Seiten 727732.

Turk, M., Collins, H.-J., Lechner, P., Heyer, K.-U., Regener, D., 1997. Ablagerungsverhalten mechanisch-biologisch behandelter Restabfalle. In: Bilitewski, B., (Hrsg): Mechanisch- biologische Verfahren zur stospezischen Abfallbeseitigung, Beihefte zu Mull und Abfall, H. 33; pp. 8799. Unger N., 2005. A review and comparison of the environmental impact of dierent materials across their life cycle. In: Cossu, R., Stegmann, R. (Eds.), Sardinia 2005 Tenth International Waste Management and Landll Symposium, 37 Oktober 2005, S. Margherita di Pula, Cagliari, Sardinien, Italien. Proceedings; CISA Environmental Sanitary Engineering Centre; pp. 907908 (Abstract; full paper on CD). Van der Ven, B., 1997. Appendix B: review of project Life cycle and solid waste. In: Sundqvist, J.-O., Finnveden, G., Stripple, H., Albertsson, A.-C., Karlsson, S., Berendson, J. (Eds.), Life Cycle Assessment and Solid Waste Stage 2. Swedish Environmental Protection Agency. Wallmann, R., 1999. Okologische Bewertung der mechanisch-biologischen Restabfallbehandlung und der Mullverbrennung auf Basis von Ener