Reviewoflcafoodwastebernstadandjansen2012.pdf

Waste Management 32 (2012) 2439–2455

Contents lists available at SciVerse ScienceDirect

Waste Management journal homepage: www.elsevier.com/locate/wasman

Review of comparative LCAs of food waste management systems – Current status and potential improvements A. Bernstad ⇑, J. la Cour Jansen Water and Environmental Engineering at the Department of Chemical Engineering, Lund University, Chemical Centre, 221 00 Lund, Sweden

a r t i c l e

i n f o

Article history: Received 31 January 2012 Accepted 21 July 2012 Available online 24 August 2012 Keywords: Life-cycle assessment Food waste System-boundaries Key issues

a b s t r a c t Twenty-five comparative cycle assessments (LCAs) addressing food waste treatment were reviewed, including the treatment alternatives landfill, thermal treatment, compost (small and large scale) and anaerobic digestion. The global warming potential related to these treatment alternatives varies largely amongst the studies. Large differences in relation to setting of system boundaries, methodological choices and variations in used input data were seen between the studies. Also, a number of internal contradictions were identified, many times resulting in biased comparisons between alternatives. Thus, noticed differences in global warming potential are not found to be a result of actual differences in the environmental impacts from studied systems, but rather to differences in the performance of the study. A number of key issues with high impact on the overall global warming potential from different treatment alternatives for food waste were identified through the use of one-way sensitivity analyses in relation to a previously performed LCA of food waste management. Assumptions related to characteristics in treated waste, losses and emissions of carbon, nutrients and other compounds during the collection, storage and pretreatment, potential energy recovery through combustion, emissions from composting, emissions from storage and land use of bio-fertilizers and chemical fertilizers and eco-profiles of substituted goods were all identified as highly relevant for the outcomes of this type of comparisons. As the use of LCA in this area is likely to increase in coming years, it is highly relevant to establish more detailed guidelines within this field in order to increase both the general quality in assessments as well as the potentials for cross-study comparisons. Ó 2012 Elsevier Ltd. All rights reserved.

1. Background 1.1. LCA of waste management Modern waste management systems are often characterized both by contribution to and avoidance of different types of environmental impacts. Energy and resources are needed initially in the management chain, while avoidance often is reached in later parts, through substitution of virgin goods. A distinction can thereby be made between direct and indirect emissions (and resource use), where the latter also can be divided into upstream and downstream emissions (modified after Gentil et al. (2010)). These processes could be described as follows:  Direct emissions, directly linked to the waste management, origin from collection/transportation, treatment and post-treatment of the waste.

⇑ Corresponding author. Tel.: +46 46 222 9557. E-mail address: [email protected] (A. Bernstad). 0956-053X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.wasman.2012.07.023

 Indirect emissions or avoided emissions (or indirect burdens or avoided burdens (Clift et al., 2000)) occurring in the background system: – Upstream indirect activities, for example production of materials and energy carriers used in the treatment chain or construction of machinery and treatment facilities used in the treatment. – Downstream indirect activities, for example avoided emissions when substituting materials and energy carriers by activities in the waste management chain. The LCA methodology provided through the ISO-standards (ISO 14001, 14041, 14044) (ISO, 2000, 2006a,b) is aimed to increase the transparency when using LCA methodology and increase the comparability between LCA-studies. However, the wording in the standards is general and does not give any detailed guidance in relation to use of LCA in specific areas, such as the management of food waste. In comparative analyses, it is vital that the system boundaries for different options are, if not equal, at least justly comparable. Assumptions and simplifications are however necessary in all LCAs. Guidance on how to manage this in the best possible way has been presented, but previous studies show that the diversity

2440

A. Bernstad, J. la Cour Jansen / Waste Management 32 (2012) 2439–2455

Table 1 Consulted guidelines related to LCA of MSW management in general and food waste in particular. Document

Waste types

Included treatment technologies

Comment

Bjarnadottir et al. (2002)

Bio-waste

ILCD (2011a)

MSW

ILCD (2011b)

Bio-waste

Sundqvist (1999)

MSW, coal ashes

Incineration, landfill, composting, anaerobic digestion, bio-cells Incineration, landfill, composting, anaerobic digestion, mechanical/biological treatment Incineration, landfill, composting, anaerobic digestion, mechanical/biological treatment Incineration, landfill, biofill, cell deposits

la Cour Jansen et al. (2007)

Bio-waste

Includes generic data on emissions from landfills, composting, anaerobic digestion and incineration Pyrolysis and gasification are described without details Gives general recommendations on general simplification rules Gives recommendations on sound allocation and system boundary setting methodologies. Includes data on landfills, composting, anaerobic digestion, incineration, MBT

PCR (2008)

MSW

Mechanical/biological treatment, incineration, composting (windrow, closed reactor, home composting), anaerobic digestion, landfill Incineration, landfill, composting, mechanical/ biological treatment

between different studies still can be large (Winkler and Bilitewski, 2007). 1.2. LCA of food waste management In contrast to other materials, such as waste paper or metal, the open loop recycling in the food waste management chain makes the discussion related to characteristics related to losses in the recycling process as well as effects on the demand for goods and services in the background system due to the recovery in the foreground system (Clift et al., 2000), less relevant. However, LCA of food waste management is still a complex field as it includes both technical and biological processes. A characteristic of food waste, different to many other waste fractions, is that food waste will be subject to biological processes during the waste management chain. These processes can both result in emissions with negative environmental impacts and affect the potentials for nutrient and energy recovery through different treatment alternatives. As these biological processes can be highly dependent on factors such as climate, rainfall, and soil profile, local conditions and time frame setting can be of importance to a larger extent compared to many other waste fractions. Finnveden et al. (2009) state that when an LCA includes forestry, agriculture, landfills and emissions to external wastewater systems the system boundaries needs to be explicitly defined, as the boundary between the technical system and the environment is not always obvious. The processes mentioned above, cited from Finnveden et al. (2009), are often included in LCA of food waste management strategies, leaving the LCA practitioner with several options on how system boundaries should be set in the specific study. Including all emissions and uses of resources connected to different treatment alternatives is in reality very difficult. Thus, the performer will include some, exclude others and use simplifications and cut-offs, depending on used system boundaries. It is unclear to which extent such differences affects the results. Several previous attempts have been made to bridge the gap between the general ISO standards and the complex questions arising when performing LCA within the waste and bio-waste field (Table 1). However, it is unclear to what extent such guidance documents have influenced the LCA-community and to what extent they succeed in address relevant issues as well as if given recommendations are similar in the different guidelines. 1.3. Aim and methodology The objective of the study was to perform a critical and systematic review of previous LCA-studies where alternative treatment/ disposal alternatives of food waste were compared with the aim of identifying and assess the most decisive system parameters and boundary assumptions for the conclusions obtained in

Includes data on energy content, biogas production and collection

reviewed studies. The aim is also to discuss how differences in system parameters and boundary assumptions can affect the result of the LCA and make recommendations with the aim of increasing both the internal comparability between treatment alternatives in a specific LCA but also between different LCAs. As a ranking of treatment alternatives in reviewed studies was assessed only in relation to GHG-emissions, a special focus was given to factors with influence on global warming potential (GWP) in the review. However, also emissions of substances with no direct impact in relation to GWP (such as NO3, NH3 and P) are discussed, as they are of relevance indirectly in relation to GWP as this means that less nutrients will be available for substitution of chemical fertilizers in biological treatment alternatives. The focus of the review is restricted to the life-cycle inventory (LCI). This limitation was made as LCIA for global warming in most methods are based on IPCC methodology, making comparisons of different LCIA methods less interesting. A review matrix was developed for the systematic review of selected papers, divided in five main treatment steps (Table 2). Based on this, system boundary settings were compared in relation to the need for consistency over systems (i.e. internal consistency) and between studies. In order to relate the outcomes from the review with previous recommendations to LCA practitioners, these were continuously compared to a selection of available guidelines (see above). Based on the outcomes from the literature review, a series of sensitivity analyses were performed in relation to a previous comparative LCA of household food waste management, based on a Swedish case study (Bernstad and la Cour Jansen, 2011). The analyses were used as a basis for a discussion of the importance of differences noticed in the review in relation to the overall outcomes of comparative LCAs.

2. Reviewed studies Four criteria were stated in the selection of studies to be included in the review. Studies should: 1. Include a thorough assessment of different treatment alternatives for food waste. 2. Include at least one biological treatment alternative for food waste. 3. State the aim to use LCA methodology, referring to ISO standards or other LCA guideline documents. 4. Include global warming potential (GWP) as one of assessed impact categories. 5. Have been published between the years 2000 and 2010. Twenty-five comparative assessments (2–5 treatment technologies) were reviewed. In several cases, studies address only food waste treatment and in some cases the study compare scenarios

2441

A. Bernstad, J. la Cour Jansen / Waste Management 32 (2012) 2439–2455 Table 2 Review matrix for the systematic review divided in five phases. Phase

To what extent were the following issues considered in assessed studies

Input material, disposal and transportation

Heating value, CH4 potential, content of N, P, K, lignin, heavy metals, organic pollutants Collection material (production and disposal) Collection/transportation

Pretreatment

Consideration of losses (water, carbon, nutrients) in collection/storage phases, prior to treatment Consideration of necessary pretreatment and further handling of residues from the pretreatment

Treatment

Environmental loads (including biogenic CO2) related to ancillary energy and material use in the treatment

Post treatment

Environmental impact from treatment/use of produced goods and solid/liquid residues (compost, digestate, biogas, ashes, leachate etc.)

Compensatory system

Environmental impacts from production and use of substituted goods Considerations taken to energy, nutrient recovery and carbon sequestration Alternative use of treatment capacity (i.e. liberation of treatment capacity in reference system when new systems are introduced)

were food waste either is co-treated with mixed MSW or as a separate fraction, using one or more biological treatment alternatives (Table 3). Selected studies only address treatment alternatives currently available on a commercial scale. Food waste is the term used in the present paper, which demands some explanation. The revised European Union WFD (Waste Frame Directive) encourages

separate collection and recycling of bio-waste, defining bio-waste as biodegradable garden and park waste, food and kitchen waste from households, restaurants, caterers and retail premises and comparable waste from food processing plants (European Parliament, 2008). This paper focuses on the latter part of this definition, excluding investigation of systems solemnly focusing on collection and

Table 3 Studies included in the present paper including information of assessed treatment alternatives, geographical scope, used LCA software tool and assessed environmental impact categories (MSW = A full MSW system was addressed, with separate biological treatment of food waste as one possible alternatives. FW = Different treatment alternatives of food waste exclusively were addressed. FW + GW = Different treatment alternatives of food waste co-collected and treated with garden waste were addressed). Nr

Technologies

Country

Waste

LCA software

Assessed impact categories

Reference

1 2 3 4

C/AD/L C/AD/I C/I /L C/L

Indonesia Denmark Turkey Italy

None ORWARE SimPro7 Simapro7

GWP, GWP, GWP, GWP,

Aye and Widjaya (2005) Baky and Eriksson (2003) Banar et al. (2009) Blengini (2008a)

5

C/AD/L

Italy

SimaPro7

GWP, ODP, AP, EP, POF, EU

Blengini (2008b)

6

Denmark

Easewaste

GWP

Boldrin et al. (2009)

7 8

C (small/large scale, subst. of peat/chemical fertilizers) L/C AD/C/I

FW FW MSW (FW) MSW (FW + GW) MSW (FW + GW) FW

Spain Sweden

MSW (FW) FW

Simapro6 None

GWP, AP, EP, POF, ADP, ODP GWP, AP, EP, POF

9 10

L/C I/AD

US Thailand

None SimaPro

GWP, AP, EP, POF, EP, HH, AP, EUI GWP, AP, EP, POF, EU

11 12 13

I/AD/L C/I/AD using FWD I/AD

Italy US Denmark

FW MSW (FW + GW) MSW (FW) FW FW

Bovea and Powell (2006) Börjesson and Berglund (2007) Cabaraban et al. (2008) Chaya and Gheewala (2007)

SPIonExcel None Easewaste

GWP, AP, EP, Dioxins GWP, AP, EU GWP, AP, EP, POF, ETP, HTP(w), HTP(s), HTP(a)

14 15 16 17

I/AD/C/L I/AD /C I/AD/C I/C/Feed-production

Spain Japan Singapore South Korea

FW FW FW FW

TRACI None None None

GWP, HH, AP, POF, ETP, LU, WU HTP, ODP GWP, AP, EP, LU, HTP GWP, AP, EP, POF, EU GWP

I/AD

Denmark

MSW (FW)

Easewaste

Kirkeby et al. (2006)

19 20

I/AD/C (small/large scale) I/Feed production/L/C

FW FW

ORWARE SimaPro7

21

C (small/large scale)

Sweden South Korea Spain

GWP, AP, EP, POF, HTP(w), HTP(s), ET(wc), ET(wa), ET(s) GWP, EU, Nutrient recovery GWP, AP, EP, HTP, ET

FW

SimaPro7

GWP, AP, EP, POF, ADP, ODP, EU

22

Denmark

FW

Easewaste

GWP

23

AD (subst. of vehicle fuel/ electricity) C/I

Martínez-Blanco et al. (2010) Møller et al. (2009)

Italy

SimaPro7

GWP, HTP, AP, POF

Rigamonti et al. (2009)

24 25

I/AD/C/L I/AD/C/Urine separation

UK Sweden

MSW (FW + GW) MSW (FW) MSW (FW)

None None

GWP GWP, AP, EP, POF, RU FEU, RFU, Phosphorus

Smith et al. (2001) Sonesson et al. (2000)

AP, EP, POF AP, EP ADP, HTP, AP, EP, POF ODP, AP, EP, POF, EU

Cherubini et al. (2009) Diggelman and Ham (2003) Fruergaard and Astrup (2011) Güereca et al. (2006) Hirai et al. (2000) Khoo et al. (2010) Kim and Kim (2010)

18

Kärrman et al. (2005) Lee et al. (2007)

FWD: Food waste disposer; AP: Acidification potential; EP: Eutrophication potential; GWP: Global warming potential; POF: Photochemical ozone formation; ODP: Ozone depletion potential; FEU: Fossil energy use; ETP: Ecotoxicity potential; REU: Renewable energy use; EU: Energy use; ADP: Abiotic depletion potential; HH: Human health; HTP: Human toxicity potential; HM: Heavy metals; HT(w): Human toxicity (water); HT(s): Human toxicity (soil); HT(a): Human toxicity (air); ET(wc): Ecotoxicity (water cronic); ET(wa): Ecotoxicity (water acute); ET(s): Ecotoxicity (soil); LU: Land use; WU: Water use; RU: Resource use.

2442

A. Bernstad, J. la Cour Jansen / Waste Management 32 (2012) 2439–2455

treatment of garden and park waste. However, as several different terms currently are used in parallel, the review also includes studies using the term organic waste. In some cases, this term includes only food waste (Kirkeby et al., 2006; Banar et al., 2009; Cherubini et al., 2009; Bovea and Powell, 2006; Sonesson et al., 2000), in other cases, the same term includes also yard and garden waste (Rigamonti et al., 2009; Chaya and Gheewala, 2007; Blengini, 2008a,b). Thus, in these four studies were food waste and garden waste was assumed to be co-collected, co-transported and co-treated were also included in the review. For the purpose of this paper, the majority of important studies are believed to be included; however, some may unintentionally have been left out. The selected studies vary greatly in relation to key assumptions and system boundary-setting, which will be described in Sections 3 and 4. Thus, the aim with the present study is not to compare studies with similar key assumptions and system boundary setting, but rather studies stating that they have the same aim, i.e. to compare the environmental impacts related to different methods for treatment of the food waste fraction in MSW. 3. Results and discussion 3.1. Comparison of GWP and ranking between alternatives A comparison between the outcomes of the studies in relation to GHG-emissions is presented in Fig. 1. It is clear that the difference between the studies is large; from net reductions to net increases of GHG-emissions in relation to all assessed treatment alternatives, with the exception of landfill where a net increase was reported in all cases. As seen in Table 4 the ranking varies largely between the reviewed studies. Incineration, AD and composting have all been ranked as the alternative most as well as the least contributing to GWP, while landfilling in all cases has been ranked as one of the least beneficial alternatives. Thus, although some coherence can be seen, variations in absolute values of GHG-emissions from respective treatment alternative as well as the ranking of different alternatives are striking. Thus, one of the main aims in the following review is to identify and discuss the reasons for these differences structured in the five phases presented in Table 2. 3.2. Input material, disposal and transportation 3.2.1. Input material The total carbon content as well as the carbon matrix (i.e. content of carbon in more and less easily biologically degradable

forms) in the treated waste is, independently of the chosen treatment alternative, related to many factors with direct and indirect impact on GHG-emissions, such as heating value, potential biogas production, methane emissions from landfills, potential peat substitution and potential carbon sequestration. The importance of in addressing the content of nutrients (N, P, K) and heavy metal is highlighted by both ILCD (2011b) and Bjarnadottir et al. (2002), as this will affect potentials for use of bio-solids as substitution for chemical fertilizers. Nevertheless, reviewed studies make very different assumptions regarding the initial carbon content and carbon matrix in treated waste and only a few of the reviewed studies present any information regarding the carbon matrix in treated waste, although the ratio between different types of carbon strongly could affect the mass-flow of carbon over the treatment chain. As an example, carbon sequestration through different types of waste treatment alternatives can be strongly correlated to the carbon matrix, and especially the content of lignin and hemicelluloses, commonly not degraded to the same extent as simple sugars (Martin and Haider, 1971). Also the nutrient content in treated waste is of relevance, and also here assumptions differ (Table 5). 3.2.2. Pre-collection and collection Household waste is commonly stored in household or waste bins during a shorter period before collection. Previous studies have shown a weight reduction in food waste collected in paper bags of more than 25% in a weekly collection scheme, mainly due to evaporation of water. Evaporation from waste collected in plastic bags was insignificant (Swedish Waste Management Association, 2010). Such losses changes the concentration of carbon, nutrients and pollutants in collected waste and changes all following environmental impacts when seen on a per ton generated waste basis. It could therefore be motivated to compensate for potential evaporation in comparative studies where different treatment alternatives can result in differences in evaporation prior to collection, for example through use of plastic bags in some cases and paper bags in others. Potential impacts from storage prior to treatment are addressed only by ILCD (2011a) in relation to emissions to air but not to water. In studies where an extra amount of waste bags needed for separate collection of food waste have been addressed, it is concluded that this can have a large impact on the environmental performance of the system (Blengini, 2008a; Bovea and Powell, 2006; Kirkeby et al., 2006). In all these cases, virgin plastic bags were used. As an example, Kirkeby et al. (2006) conclude that the use of plastic bags used for food waste collection is decisive in ranking of thermal treatment as superior in relation

Fig. 1. GWP from 1 ton of food waste treated with different technologies according to reviewed studies (see Table 1 for reference to respective study).

2443

A. Bernstad, J. la Cour Jansen / Waste Management 32 (2012) 2439–2455

Table 4 Ranking of the treatment alternatives in relation to the environmental impact category GWP in included studies. Incineration refers to mass-burning with energy recovery in all cases. Treatment alternative

Incineration

Number rankings as most beneficial Number rankings as least beneficial

Anaerobic digestion

Compost

No energy recovery

Energy recovery

Sanitary

Landfill Open dump

Large scale

Large scale

Small scale

0 out of 2 1 out of 2

3 out of 9 2 out of 9

0 out of 7 6 out of 7

0 out of 4 3 out of 4

7 out of 11 2 out of 11

4 out of 12 2 out of 12

3 out of 8 1 out of 8

Table 5 Nutrient content in treated food waste in reviewed studies. kg/ton wet waste

N

P

K

C-tot

Lignin

Range Average Standard deviation Number of studies

1.4–14.1 6.7 4.2 7

0.7–2.0 1.2 0.5 4

2.8–3.0 2.9 – 1

83–365 145 34 12

3–12 8 4 4

to anaerobic treatment and Blengini (2008a) state that production and delivery of plastic bags used for separate collection of food waste represented 55% and 33% of the total energy consumption and photochemical ozone formation in the assessed food waste recycling scenario. Use of specific data for material composition of the litter bins/containers and of the plastic bags is urged by PCR (2008) while not mentioned in other consulted guidelines. 3.2.3. Transportation It should be remembered that some alternatives (i.e. home composting) will not include any need for collection and transportation. However, several previous studies have shown that environmental impacts related to transport of waste seldom is of large importance to the overall environmental impacts of the investigated system (Tyskeng and Finnveden, 2007; Tan and Khoo, 2006). This is the conclusion also in the majority of the reviewed papers and transports were in some cases excluded from the LCI with reference to previous indications of the low importance of such processes in similar studies (Chaya and Gheewala, 2007; Khoo et al., 2010). However, consulted guidelines address collection and transportation of waste as important parts of the LCI and commonly reserve extensive space for considerations on how such should be addressed. The only exception is found in Bjarnadottir et al. (2002), where impacts from transportation are seen as irrelevant. 3.3. Pretreatment 3.3.1. Physical pretreatment Some sort of physical pretreatment of food waste is often needed before biological treatment can be accomplished. However, energy and resource input connected to pretreatment is seldom taken into consideration in reviewed studies. Energy use in pretreatment of food waste has previously been reported in levels of 8–80 kW h/ton incoming food waste (Bernstad and la Cour Jansen, 2011). Investigations of modern pretreatment plants in Denmark and Sweden have also shown that the mass-losses due to generation of rejects in the pretreatment can reach 2–45% of incoming food waste depending on the chosen technology (Hansen et al., 2007; Truedsson, 2010). In cases where the amount of reject generated is large, the effect of pretreatment in relation to the massflow through the treatment process is therefore probably of larger relevance to the environmental impacts than energy and resource used in the process. If large parts of the potentially biologically treatable material are routed to rejects, subjected to thermal treatment or landfill, it can be argued that this should be taken into con-

sideration in the calculations of the potential energy and fertilizer substitution from biological treatment alternatives. Potential energy recovery, energy use and emissions related to further treatment of refused material should in that case also be considered. In order to determine the full environmental effects from physical pretreatment, losses of nutrients, readily degradable organic matter, slowly degradable organic matter as well as potential impurities and toxic compounds would have to be assessed.1 A need for pretreatment of separately collected food waste is mentioned in less than 50% (10 studies) of reviewed studies. However, even fewer actually take pretreatment into consideration in the LCI. Smith et al. (2001) take material losses in pretreatment into consideration in relation to composting but not to AD, while Sonesson et al. (2000) do the opposite (Table 6). In both cases, these inconsistencies increase the risk for biased comparisons between treatment alternatives. In most cases (six out of ten studies) further treatment of rejects is not addressed resulting in a non-finalized material flow of rejects and a risk for biased comparisons between different treatment alternatives. In some cases, rejects are incinerated with energy recovery, having a large positive impact on the overall results from the study due to the substitution of other energy carriers (Baky and Eriksson, 2003; Kärrman et al., 2005). la Cour Jansen et al. (2007) emphasises the importance of addressing physical pretreatment, both in relation to energy use in the process as well as distribution factors between refuse and bio-fractions. Bjarnadottir et al. (2002) mention a potential need for pretreatment, but do not give any recommendations on how this should be handles, while other guidelines included in this review do not do not take the need for physical pretreatment prior to biologic treatment into consideration.

3.4. Treatment 3.4.1. Direct emissions from landfills Emissions from landfills will in reality occur over very long time periods. However, restricted timeframes are often used in LCAs. Choices in relation to timeframe setting can be of large importance in relation to GHG-emissions from landfills, but both guideline recommendations and values used in reviewed studies differ largely (Table 7). The decay rate in landfilled waste is according to the IPCC methodology dependent both on the carbon–matrix in landfilled materials as well as geographical factors (IPCC, 2006). However, a majority of reviewed studies refer to literature values for accounting of CO2 and CH4 emissions, without any further discussion of a need to adjust these in relation neither to the carbon content in landfilled waste nor to geographical differences. Emissions of carbon through leachate are rarely addressed, but when this is the case, such emissions are assumed to represent only 1% of total carbon losses (Sundqvist, 1999; Aye and Widjaya, 2005). Emissions of N2O in landfill gas can according to ILCD (2011a) be left unconsidered, which was the case in all reviewed studies. 1 The same logic can be applied to separation of contaminations post-treatment, for example through sieving of compost. In some cases it is not clear whether residues emerge before or after treatment and in the present paper these are discussed together.

2444

A. Bernstad, J. la Cour Jansen / Waste Management 32 (2012) 2439–2455

Table 6 Material losses, energy use and treatment of residues from pre-treatment of food waste in reviewed studies. Reject level (% wet waste)

1 2 3

Composting

AD

6–171

13–352

Energy use (kW h/ton)

Treatment of residues

Energy recovery from incineration of residues (MW h/ton)

0–25

Incineration, landfill, site restoration

0.7–1.5 (heat)+0.4(electricity)3

Bovea and Powell (2006), Baky and Eriksson (2003), Banar et al. (2009), Khoo et al. (2010), Kärrman et al. (2005), Lee et al. (2007), Smith et al. (2001), Kim and Kim (2010). Baky and Eriksson (2003), Kirkeby et al. (2006), Kärrman et al. (2005), Sonesson et al. (2000). Baky and Eriksson (2003), Kärrman et al. (2005).

Table 7 Direct emissions from landfill of food waste addressed in reviewed studies as kg CH4/ton ww. CH4 (kg/ton)

Methodological approach

Timeframe (years)

31–761,2,3 56 52

Use of literature data 100% degradation of DDOC4 in waste, minus oxidation factor (0.1) 95% degradation of bio-available carbon, of which 84% during the design lifetime

501, 1002, infinite3 100 15 (design time)

85 99

Calculated based on C, H,O, N, S content (according to Finnveden, 1992; Sundqvist, 1999) 100% degradation of bio-available carbon minus carbon lost from the landfill through leakage n.s.

100 100 years or infinite

n.s. 77 n.s. 1 2 3 4

Use of literature data 100% degradation of proteins, sugars, starch, 70% degradation of cellulose. Oxidation factor = 0.1–0.15. Leachate factor 0.01

100 years and long-term (separately) 30 Pseudo steady-state (100 years) and infinite

Reference Smith et al. (2001) Diggelman and Ham (2003) Kim and Kim (2010) Bjarnadottir et al. (2002) ILCD (2011a) PCR (2008) Sundqvist (1999)

Aye and Widjaya (2005). Blengini (2008a,b), Cherubini et al. (2009), Güereca et al. (2006). Lee et al. (2007). Dissimiable degradable organic carbon (assumed to be 75% of ww in food waste by Smith et al. (2001)).

3.4.2. Direct emissions and energy recovery from incineration There is little agreement on what types of emissions that should be taken into consideration in case of incineration of food waste and the relevance of made choices are of course strongly linked to the environmental impacts taken into consideration in the study. Bjarnadottir et al. (2002) suggest that pollutants regulated in the EU incineration of waste directive should be taken into consideration. However, in relation to GHG-emissions it is seen that both CH4 and N2O, although not addressed by the EU directive mentioned above, should be addressed according to la Cour Jansen et al. (2007). PCR (2008) gives no guidelines on emissions from incineration. Direct emissions from food waste incineration can be divided into input and process specific emissions, i.e. related to the content of C, N, S, Cl, F and metals possibly present in combusted material or dependent on the process conditions in the specific incineration plant (i.e. flue gas treatment, treatment of ashes etc.). However, emissions from incineration scenarios presented in reviewed studies are commonly based on literature data, and many times it is unclear if considerations are taken to differences in input material as well as in combustion technologies. A remarkable diversity in the assumed energy recovery from combustion of food waste was seen (Table 8). However, in some studies it was also assumed that combustion of food waste should be credited for no net energy recovery at all (Chaya and Gheewala, 2007; Diggelman and Ham, 2003) due to the high moisture content in food waste. Thus, combustion of food waste can result in significant GHG-emissions due to the need for support fuel, although emissions related to support fuel combustion are not always addressed. An explanation for some of these differences could be related to utilized combustion technologies, where a large part of the energy needed for evaporation of water in combusted food waste can be recovered in plants with flue gas condensation, resulting in overall higher energy recovery. There is also a need for energy input in flue gas cleaning processes etc., which in many cases is not considered (Table 8). In several

studies, no consideration was taken to further treatment of slag and ashes from combustion of food waste (Banar et al., 2009; Güereca et al., 2006; Khoo et al., 2010; Rigamonti et al., 2009). Energy use and emissions related to such further treatment processes should according to Bjarnadottir et al. (2002), PCR (2008) and Sundqvist (1999) be addressed, while they are seen as insignificant by la Cour Jansen et al. (2007). Use of auxiliary materials (lime, chemicals etc.) for flue gas cleaning etc. is with few exceptions (Fruergaard and Astrup, 2011) omitted, which is also recommended by la Cour Jansen et al. (2007). 3.4.3. Direct emissions from anaerobic digestion Fugitive emissions of CH4 from anaerobic digestion plants may occur during the AD-process. Such emissions vary between different facilities and can according to Eggleston et al. (2006) range between 0 and 10 vol.% of produced biogas, but are according to the same author likely to be closer to 0% in plants where unintentional emissions are flared. Previous assessments of full-scale plants have shown that fugitive emissions of CH4 can vary largely (Swedish Waste Management Association, 2007). The IPCC guidelines assume fugitive emissions of methane from biogas production plants to 1 g CH4/kg waste (dry weight) with a range of 0–8 g (IPCC, 2006). This would be equivalent to an emission of 1.4 vol.% (range 0–11 vol.%) with a methane production of 100 Nm3. Fugitive methane from AD plants were assumed to be 0 or 0.5–2% of produced methane (0.5–1.2 kg CH4/ton waste) in reviewed studies. While Møller et al. (2009) and Börjesson and Berglund (2007) state that fugitive emission of methane is a key parameter in relation to GHG-emissions from AD systems (in both cases assuming emissions equal to 0–3 vol.% of produced methane), others (Aye and Widjaya, 2005; Diggelman and Ham, 2003; Chaya and Gheewala, 2007; Cherubini et al., 2009; Güereca et al., 2006) do not take such emissions into consideration at all in their assessments. Fugitive methane emissions are seen as relevant by both Bjarnadottir et al. (2002) and la Cour Jansen et al. (2007) (Table 9). Direct energy

2445

A. Bernstad, J. la Cour Jansen / Waste Management 32 (2012) 2439–2455

Table 8 Summary of assumptions regarding energy content in food waste, energy recovery and energy input in incineration processes as well as generation of residues (n.s. = not stated, n.c. = not considered). LHV (GJ/ton wet waste)

LHV (GJ/ton dry waste)

1.7–6.31

8.6–21.0

6.98

19.78

Energy production (kW h/ton) Electric

CO2 (kg/ ton)

N2O (kg/ ton)

Energy input (kW h/ ton)

Ash/slag generation (% of wet waste)

Reference

453–8224

0.01–0.155

70–5336

0–25%7

200–8809S

n.s.

8010

14.6%8

Reviewed studies Guidelines

Thermal

89– 553– 5202 13643 480 (total)8

1

Aye and Widjaya (2005), Baky and Eriksson (2003), Börjesson and Berglund (2007), Fruegaard and Astrup (2011), Smith et al. (2001), Lee et al. (2007), Güereca et al. (2006). 2 Baky and Eriksson (2003), Fruegaard and Astrup (2011), Kärrman et al. (2005), Smith et al. (2001). 3 Baky and Eriksson (2003), Börjesson and Berglund (2007), Fruegaard and Astrup (2011), Kärrman et al. (2005), Smith et al. (2001), Khoo et al. (2010), Lee et al. (2007). 4 Baky and Eriksson (2003), Khoo et al. (2010), Lee et al. (2007), Smith et al. (2001). 5 Baky and Eriksson (2003), Khoo et al. (2010), Smith et al. (2001). 6 Aye and Widjaya (2005), Diggelman and Ham (2003), Güereca et al. (2006), Fruegaard and Astrup (2011). 7 Baky and Eriksson (2003), Diggelman and Ham (2003), Güereca et al. (2006), Kärrman et al. (2005), Khoo et al. (2010), Lee et al. (2007), Smith et al. (2001). 8 Sundqvist (1999). 9 Sundqvist (1999), Bjarnadottir et al. (2002), la Cour Jansen et al. (2007). 10 la Cour Jansen et al. (2007).

Table 9 Summarized assumptions of emissions from biological treatment plants, storage, transportation and spreading of fertilizers and fertilizers from farmland as well as replacement of chemical fertilizers in the studies and guidelines. N2O(d) = Direct N2O-emissions, N2O(i) = Indirect N2O-emissions, HM = Heavy metals. Numbers of studies where a specific emission was considered out of total number of studies in each treatment type is showed within brackets. Type

Treatment

Compost

CO2 (1/19) CH4 (5/19) NH3 (8/19) N2O (4/19) VOC (1/19) COD/BOD (3/19) NH4+ (2/19) CH4 (10/16)

Anaerobic digestion

Storage

CH4 (1/16) NH3 (1/16) N2O(d) (1/16) N2O(i)5 (0/16)

Transports/spreading

Farmland emissions

Substituted chemical fertilizers

Substitution ratio

Yes (5/19)

NH3 (1/19) N2O (1/19) NO3 (1/19) HM (1/19)

N (12/19) P (10/19) K (8/19)

0%1 (N, P, K) 20–100% (N) 0–100% (P,K)

Yes (3/ 16)

NH3 (2/16) N2O (2/ 16) 2 NO3 (1/ 16) HM (2/16) NO3 (3/25) N2O(d)2 (0/25) N2O(i)2 (0/25)

N (11/16) P (9/16) K (8/16)

40–100% (N) 0–100% (P,K)

Chemical fertilizers

1 2

Only considered when compost substitutes chemical fertilizers and not when substituting peat by Boldrin et al. (2009). Addressed by IPCC (2006).

use in AD plants varies largely and should according to all assessed guidelines be addressed.

3.4.4. Direct emissions from composting Energy input in composting varies largely in reviewed studies; 15.1–55.0 kW h electricity/ton ww and/or 0.01–15.3 L diesel/ton ww. Emissions of both carbon and nitrogen containing compounds take place during aerobic degradation of food waste. The form and ratio of these emissions will to a large extent depend on factors such as moisture content, oxygen supply and C/N-ratio during the composting process (Kirchmann, 1985). Such emissions are addressed to a varying extent in reviewed papers (Table 9). An absolute majority of the studies consider biogenic CO2-emissions from composting as neutral in relation to GWP, only in three studies where biogenic emissions from treatment of food waste attributed with a GWP factor of 1 (Blengini, 2008a,b; Lee et al., 2007). As concluded by Christensen et al. (2009), biogenic CO2 emissions can be seen both as neutral or contributing to GWP, as long as a consistency is made both throughout a specific system and between compared systems. However, in all cases where biogenic emissions from treatment of food waste attributed with a GWP factor of 1, emissions of CO2 from later use on-land of produced bio-fertilizers were seen as neutral in relation to GWP. This disfavours incineration and landfill alternatives in relation to biological treatment

alternatives. In studies where emissions of N2O and CH4 from composting were taken into consideration, authors in many cases conclude that such emissions can have a large effect on the overall GWP from these systems (Boldrin et al., 2009; Sonesson et al., 2000; Börjesson and Berglund, 2007). In studies where NH3-emissions were taken into consideration it was seen that these in many cases have a large impact both on acidification and eutrophication. Guidelines generally recommend emissions of CO2, N2O, CH4 and NH3 to be considered, although ILCD (2011a) argue that toxic emissions other than CO2 should be excluded from closed composting method plants due to efficient filtering of air emissions. Emissions to water are considered in few studies (Aye and Widjaya, 2005; Bovea and Powell, 2006; Kim and Kim, 2010; Cabaraban et al., 2008) while others, in line with Bjarnadottir et al. (2002) assume that production of leachate and emissions related to this either do not occur or are insignificant. However, emissions through leachate formation are seen as relevant in two of the consulted guidelines (ILCD, 2011a,b).

3.5. Post treatment 3.5.1. Use of organic residual fractions The assessment of environmental impacts connected to further treatment or use of compost and digestate can vary largely. In

2446

A. Bernstad, J. la Cour Jansen / Waste Management 32 (2012) 2439–2455

Table 10 Carbon emissions over the treatment chain, as kg C/ton food waste, as reported in the five reviewed studies were the initial carbon content was presented (n.r. = not relevant for this type of treatment alternative, n.a. = carbon emissions during this process could occur but were not addressed, 0 = the process has been considered but no emissions of carbon are assumed to occur, C = Compost, I = Incineration, L = Landfill, AD = Anaerobic Digestion). Treatment

1 2

Upgrading

CO2

CH4

98 0 130 12 113 12–113 63 191 13.9 158 57 0 0 0 112.5 112.5 147 56.3

0 1 0 0.02 5.1 0.02–5.1 64 64 38.1 0 0 0 0.5 2.8 0 0 0 56.3

n.r. 01 n.r. n.r. n.r. n.r. n.r. n.r. n.r. n.r. n.r. 0.41 0.61 n.r. n.r. n.r. n.r. n.r.

Storage

0 0 n.r. 0 0 0 n.r. n.a. n.r. n.r. n.a 0 0 n.a. n.a. n.a. n.r. n.r.

Use on-land

0 0 n.r. 55 332 25–154 n.r. 0 n.r. n.r. n.a. 43.2 75.7 0 68.2 40.3 n.r. n.r.

Use of gas CO2

CH4

n.r. 97 n.r. n.r. n.r. n.r. 0 n.r. 0 n.r. n.r. 42 68 0 n.r. n.r. n.r. 0

n.r. 01–1.22 n.r. n.r. n.r. n.r. 0 n.r. 0 n.r. n.r. 11.32/0.41 18.02/0.6 31.8 n.r. n.r. n.r. 0

C-sink

C-tot initial

Non-accounted

Reference

7.6 7.6 0 0.5 21.5 0 0 0 0 0 0 1.8 12.3 16.5 0.24 0 0 36.8

134.5 134.5 134.5 100 365 100–365 255 255 158 158 158 113 130 190 190 190 190 190

29 28 4.4 33 106 63–93 129 64 104 105 101 58–97 28–114 142 83 39 43 41

Baky and Eriksson (2003) C Baky and Eriksson (2003) AD Baky and Eriksson (2003) I Boldrin et al. (2009) C (low) Boldrin et al. (2009) C (high) Boldrin et al. (2009) C (peat subst.) Kim and Kim (2010) L Kim and Kim (2010) C Lee et al. (2007) L Lee et al. (2007) I Lee et al. (2007) C Møller et al. (2009) AD (low) Møller et al. (2009) AD (high) Smith et al. (2001) AD Smith et al. (2001) C Smith et al. (2001) C (peat subst) Smith et al. (2001) I Smith et al. (2001) L

Biogas used as vehicle fuel. Biogas used for electricity/heat production.

some cases, use of compost and digestate are connected to large environmental benefits (Møller et al., 2009; Boldrin et al., 2009; Smith et al., 2001), while the decreased environmental burden related to use of organic fertilizers is assumed to be much less relevant in some cases (Aya and Widjaya, 2005; Rigamonti et al., 2009) and in some cases it is not assumed that the quality of produced bio-fertilizers are adequate for use as fertilizers and further treatment therefore result in net burdens (Hirai et al., 2000). Several factors have been identified which could explain these differences; amount of generated bio-fertilizers (ranging from 0.013 to 1.07 ton wet compost or digestate per ton treated wet waste), emissions during storage, assumed substitution ratio in relation to chemical fertilizers, the eco-profile (the environmental burdens) of the substituted chemical fertilizers, assumed carbon sequestration and emissions from organic and chemical fertilizers when spread on farmland. Most studies do not take losses of carbon and nitrogen during storage of bio-fertilizers into consideration. However, when taken into consideration, methodologies and assumptions vary largely (Table 9). As an example Kärrman et al. (2005) take direct emissions of N2O into consideration (0.1% of N-tot or 0.007 g N2O/ton wet waste) while IPCC (2006) state that emissions of N2O from storage of organic fertilizers are insignificant, but include indirect emissions of N2O resulting from emissions of NH3 (0.5% of total N-content). Previous studies have shown that emissions of greenhouse gases from stored bio-fertilizers are strongly temperature dependent, decreasing by more than 40% when storage temperatures are changed from 20 to 9 °C (Clemens et al., 2006). Several guidelines recognize that emissions of NO3 , NH3 and N2O from fertilizers should be considered in LCAs including agro- and forestry systems (ILCD, 2010, 2011a; la Cour Jansen et al., 2007). Emissions from farmland application of organic fertilizers are taken into consideration in very different ratios in the studies. Several authors report that current knowledge is insufficient, especially in relation to N2O emissions from organic fertilizers. According to IPCC guidelines emissions also occur when chemical fertilizers are used on farmland (IPCC, 2006). Only in one case (Møller et al., 2009), net emissions from use of organic fertilizers in comparison to chemical are taken into consideration. Also the substitution ratios of fertilizers differ largely between studies (Table 9). Neither any of the studies, nor any of the consulted

guidelines take emissions of phosphorus to surface and ground water into consideration. These emissions are strongly depending on mechanisms that are not well known and the emissions can therefore not be directly related to the phosphorus in the applied fertilizers (la Cour Jansen et al., 2007). 3.5.2. Use of biogas and LFG Production of biogas from AD is reported in the range of 92– 165 Nm3/ton treated food waste, with the energy content to 2450–3890 MJ/ton (Baky and Eriksson, 2003; Fruergaard and Astrup, 2011; Smith et al., 2001; Kärrman et al., 2005; Kirkeby et al., 2006). Biogas and LFG produced from food waste can be used for energy recovery through direct combustion (production of electricity and heat) or as vehicle fuel. If the latter is assumed, produced biogas needs to be upgraded, compressed/liquefied and distributed. Previous studies have shown that both energy-demand, need for additives and fugitive emissions of CH4 during upgrading of biogas could vary largely between different upgrading techniques (Benjaminsson, 2006). Lantz et al. (2009) and Pertl et al. (2010) show that the choice of upgrading technology and subsequent assumptions regarding CH4-emissions during upgrading can have a large impact on the overall GWP from the treatment chain. According to the IPCC guidelines, combustion of recovered biogas is not significant in relation to global warming as the CO2 emissions are of biogenic origin (and thus GWP neutral) and the CH4 and N2O emissions are low (IPCC, 2006) while la Cour Jansen et al. (2007) state that emissions from combustion of biogas should be taken into consideration. Biogas combustion emissions are dependent on the type of gas engine, and lean-burn gas engines, commonly used at AD facilities, often give higher emissions of CH4 compared to other gas engines (Nielsen et al., 2008). Emissions for combustion of biogas in the papers differ between 1 and 4 g CO2-eq/kW h when used for production of electricity and heat (Baky and Eriksson, 2003; Kirkeby et al., 2006) and 0–0.6 g CO2eq/kW h when biogas is used as fuel in vehicles (Baky and Eriksson, 2003; Fruergaard and Astrup, 2011). As seen in Table 10 emissions from biogas and LFG combustion are in some cases considered when origin from AD but not from landfill, or when biogas is used for electricity/heat production, but not as vehicle fuel, which causes biased comparisons between different alternatives. It

A. Bernstad, J. la Cour Jansen / Waste Management 32 (2012) 2439–2455

should also be remembered that substitution of fuel not always can be performed on a mere energy quantity basis, i.e. 1 MJ CH4 substituting 1 MJ petrol/diesel. This is due to the fact that the combustion technique is differently developed in different engines (diesel, petrol or gas), making the driving distance of 1 MJ diesel longer than 1 MJ CH4 (Lantz et al., 2009). 3.5.3. Substituted goods The LCI considers not only direct emissions and resource use, but also indirect. The latter can be divided into upstream and downstream activities, where downstream focuses on avoided environmental impacts when goods attained in the waste treatment chain are substituting other resources, typically located outside the waste management system itself. The goods substituted through food waste treatment in the studies are electricity, thermal energy, transportation fuel, peat, chemical fertilizers and animal feed. Both the ratio of which 1 ton of treated food waste can substitute these goods as well as the environmental impacts related to both the production and use phase of these substituted goods vary largely. The most common assumption is that energy used and substituted by the assessed treatment alternative can be represented by the eco-profile of national or regional average electricity production. In some cases this approach is supplemented with a sensitivity analysis assuming a marginal electricity production. This is in line with an approach suggested by Mathiesen et al. (2009), stating that sensitivity analyses can be performed with different types of marginal energy; one fossil-lean and one fossil-intensive as a simplified strategy when and energy system analyses taking into account fundamentally different future scenarios not is applied. Such analyses typically show that the choice of electricity substituted by the waste management system can have a large impact on the overall LCA outcome (Rigamonti et al., 2009; Smith et al., 2001; Møller et al., 2009; Sonesson et al., 2000). However, in treatment systems with low need for electricity input and where no electricity is generated, the relevance of the chosen electricity eco-profile is much lower (Boldrin et al., 2009). Only in a few cases where assessed treatment alternatives render a possible thermal energy recovery is the produced heat assumed to replace other thermal energy production (Smith et al., 2001 (in the case of incineration and AD, but not in the case of landfill with LFG recovery), Sonesson et al., 2000; Møller et al., 2009; Rigamonti et al., 2009). Thermal energy is in these studies assumed to replace heat generated from such varied sources as natural gas in household boilers, wood chips, coal or an EU average heat mix. The environmental impact from the substitution is likely to be strongly connected with the assumed environmental impact connected to the substituted thermal energy. However, only in one study, a sensitivity analysis addressing the importance of this was made (Sonesson et al., 2000). Also the environmental gains from substitution of chemical fertilizers by organic is closely connected to the eco-profile of the substituted products. Emissions of N2O from Nfertilizer production are currently being reduced due to catalytic N2O reduction and reduced energy use. Emissions of N2O from fertilizer producers using best available technique (BAT) can be 64% lower and the energy demand 25% lower compared to average production plants (Jensen and Kongshaug, 2003). Based on above given examples, it is stressed that addressing the influence of technological developments affecting the eco-profile of produced goods or substitution ratio of the same through sensitivity analyses is of relevance. Another aspect that could be seen as a part of the processes substituted by the assessed system is the effect of dynamics between different treatment alternatives. As an example, separate collection and biological treatment of food waste can liberate treatment capacity in landfills or incineration plants. This can

2447

result in an extended lifetime of landfills or an increased energy production through waste incineration, assuming a higher energy value in residual waste with lesser food waste. Such dynamic effects were not taken into consideration in any of the reviewed studies. 3.5.4. Carbon sequestration and carbon mass-flow Food waste consists of both readily available carbon and carbon with a longer degradation process. Thus a fraction of the carbon content in treated wastes is bound in soil for longer periods. This carbon can be seen as a sink of CO2, credited as an avoided downstream emission of CO2. The concept of carbon sequestration is still an area of discussion where different approaches are seen in the papers. Møller et al. (2009) assume that the total amount of carbon in the incoming waste, minus CO2 and CH4 from combustion of produced biogas, times an emission storage factor of 0.14–0.04 is to be regarded as a sink when digestate is used on farmland. This neglects emissions of carbon as CO, CH4 and VOC during biogas combustion (0.21 g C/MJ according to Baky and Eriksson (2003) equivalent to 0.7 kg C/ton food waste) and during storage of digestate (8 kg C/ton food waste according to Kärrman et al., 2005). Boldrin et al. (2009) use average data on the C-tot content in compost combined with a literature data storage factor of 0.14–0.02, while Blengini (2008a) and Blengini (2008b) use an average for carbon storage based on Linzner and Mostbauer (2005), resulting in a storage factor of 0.017–0.06 for digestate and 0.036–0.131 for compost. Smith et al. (2001) assume that the degradable carbon content in food waste minus dissimiable organic carbon will be left in soil as a carbon stock when food waste is landfilled. In the case of compost and digestate, a turnover time of 42 years is used, resulting in a storage factor of 0.08. Other previous studies base the assumed emission storage factor on the lignin content in treated waste, as lignin is assumed not to degrade in the AD process, and an assumed degradation ratio of 60–85% for carbon hydrates, proteins and fat (Lantz et al., 2009). Assumptions in the studies range from 0% to 10.5% of the initial carbon content for compost, 0–4.8% for AD and 0–25% for landfill – in all cases related to a 100 year timeframe. la Cour Jansen et al. (2007) assume that roughly 10% of initial carbon content left after 100 years, based on modeling with the Daisy software (Bruun et al., 2006). ILCD (2011a) recommend LCA-practitioners not to use carbon sequestration as default, while Bjarnadottir et al. (2002) gives no absolute recommendations to whether a carbon sink approach should be used or not. In Table 10, data on carbon storage is combined with data on carbon emissions from other parts of the treatment chain and on the initial carbon content in treated food waste, which only was given in five of the reviewed studies. All emissions were re-calculated to C/ton wet waste. In cases were a ratio of data was presented, maximum and minimum levels are presented. The summary shows that the assumed carbon flows during different processes in the treatment chain can differ largely between and within compared studies. This does not necessarily mean biased comparisons between scenarios, as biogenic CO2-emissions in many cases are assumed not to contribute to GWP. However, it clearly shows that emissions assumed to occur over the treatment chain commonly are not cross-checked with substance-flow analyses. 3.6. Use of sensitivity analyses Saltelli et al. (1995) defines sensitivity analyses as the contribution of the uncertainty in individual input parameters to the predicted model result. However, the term sensitivity analysis is commonly used both to assess the robustness of the results gained in the study in relation to uncertainties in used input data (Smith

2448

A. Bernstad, J. la Cour Jansen / Waste Management 32 (2012) 2439–2455

Household sorting efficiency High (>20%)

Carbon sequestration

Chemical fertilizer environmental profile

Medium (20-10%)

Substitution factor (chemical fertilizers)

Low (<10%)

Transports Timeframe setting Post-treatment (handling of ashes from incineration) Collection material Storage of organic fertilizers Emissions from organic fertilizers on farmland Energy recovery ratio Input material quality Treatment emissions Substituted energy 0

2

4

6

8

10

12

14

16

Number of performed sensitivity analyses in respective study Fig. 2. Sensitivity analyses and outcomes. The terms High, Medium and Low are used to illustrate a >20%, 5–20% and 0–5% change of results in relation to the reference results in the different studies.

et al., 2001; Börjesson and Berglund, 2007) and to assess the effect from alterations of the monitored systems (Kirkeby et al., 2006; Rigamonti et al., 2009; Baky and Eriksson, 2003), although the latter is similar to what is called scenario analyses or scenario alterations in other studies (Koning et al., 2010). The parameter most commonly assessed through sensitivity analyses was related to electricity generation. In cases where the treatment process resulted in a net energy production and thus substitution of other energy, the eco-profile of substituted energy has a large impact on the overall results from the study, principally in relation to GWP. However, in relation to treatment scenarios where energy was addressed only as an input, as in composting (Boldrin et al., 2009) or when differences in the eco-profile related to substituted energy matrixes was small (Baky and Eriksson, 2003), the impact tended to be lower (Fig. 2). 4. Key parameter identification through performance of sensitivity analyses The relative importance of differences in system boundary setting, input data and assumptions were investigated in relation to a previous comparative LCA of household food waste management comparing incineration, composting and AD (Bernstad and la Cour Jansen, 2011), further on referred to as the reference study. Landfill was not included in the study, as there is a common view in reviewed papers that landfill is the least beneficial treatment alternative for food waste in relation to GWP (only one exception to this was seen) and as the present paper is written with the background of a need to reduce the amount of biodegradable waste disposed of in landfills (in accordance of the EU Landfill Directive). In the reference study, it was seen that AD with use of biogas and digestate as substitution for vehicle fuel and chemical fertilizers (N, P and K) respectively, results in greater avoidance of GWP and POF compared to composting or incineration of food waste, while both anaerobic and aerobic biological treatments increase net contribution to EP and AP compared to incineration. Several different methodologies have been suggested in addressing of uncertainties in LCA-performance (Björklund, 2002; Clavreul and Christensen, 2011). The methodology chosen in the present study builds on a methodology for identification and testing of key parameters presented by Eriksson and Baky (2010) and is based on performance of one-way sensitivity analysis. One-way sensitivity analysis determines the amount an individual input parameter value, all other parameters held constant, needs to

change in order for output parameter values to change by a certain percentage (USEPA, 1995). Through this method, sensitivity ratios are defined for each parameter as the ratio between the relative change in the result in a specific impact category and the relative change in a parameter. The methodology was divided in three steps: (1) Identification of activities of interest for performance of sensitivity analyses. Four areas where considered to be of interest in the identification, based on the literature review: – Activities reported to have a large impact on direct emissions, energy recovery, energy or resource use for a specific process or on the overall results – Activities reported as being connected with large uncertainties – Activities where values in reviewed studies vary largely – Activities which are included as a part of the treatment chain in some studies but not in others and where the importance of this difference is unknown (2) Performance of sensitivity analyses (3) One-way sensitivity analyzes can be applied with either arbitrarily selected ranges of variation or variations representing known ranges of uncertainty, also known as uncertainty importance analysis (Björklund, 2002). In the present study, a mixture is used, as ranges of uncertainty are known in some cases and not in other. Sensitivity ratios were identified by altering chosen input parameters in order to give a 20% change in output values in outcomes from the reference study. The needed relative change in the specific parameter was denoted. Also, to shed more light on the discussion on significant/insignificant processes and setting of system boundaries, the effect of excluding or including whole processes (such as collection, use of collection material, pretreatment or storage of organic fertilizers) were addressed. As landfilling of food waste was not included in the reference study, key-issue identification was not applied to this treatment alternative. (4) Evaluation of results from sensitivity analyses in relation to results from the reference study (5) After performance of sensitivity analyses, an evaluation process is needed where one reflects on whether the outcomes of the analyses are to be seen as relevant or not. As pointed out by Heijungs and Huijbregts (2004) there is a common confusion between usages of the terms significant and large

A. Bernstad, J. la Cour Jansen / Waste Management 32 (2012) 2439–2455

2449

Table 11 Summary of results from sensitivity analyses; relative impacts related to effects from system boundary setting changes as%-change in relation to GWP from different treatment alternatives investigated in the reference study (AD = Anaerobic digestion, I = Incineration, C = Compost). Process

% change (GWP)1

Treatment alternative

Comment

14–19 1–10 <5 30

All All AD AD

Assuming plastic bags instead of paper bags

330

I

Swedish power mix was substituted in the base case

690

AD

Swedish power mix was substituted in the base case

C AD AD I I AD AD I AD AD C

Swedish power mix was substituted in the base case Recovered energy substituting Danish coal power Recovered energy substituting Swedish average power Assuming use of energy from Danish coal power Assuming use of Swedish average power Assuming use of energy from Danish coal power Assuming use of Swedish average power

Excluding emissions of N2O Excluding fugitive emissions of CH4 (digestion) Excluding fugitive emissions of CH4 (upgrading) Including emission control (NH3 and CH4)

320 22 <5 25 <5 15 <5 <5 8 12 25

Including emissions to water Excluding use of auxiliary material Excluding use of auxiliary material Including use of auxiliary material

<5 <5 <5 55

C C I I

Assuming 1% emission Assuming 1% emission Reduction with 95% (NH3) and 25% (CH4) according to Liang et al. (2000), Busca and Pistarino (2003), and Dalemo et al. (1997) Assuming emissions according to Aye and Widjaya (2005) Not including use of wood pellets in composts Not including use of lime and urea in flue gas treatment Assuming use of oil according to Lee et al. (2007)

29 12

AD AD

Including storage with assumptions from Lantz et al. (2009). Including storage with assumptions from Lantz et al. (2009).

5–26

C

Using carbon sequestration factors from Smith et al. (2001) (0.08) and Boldrin et al. (2009) (0.14)

2–17 <5

AD I

32

C

40 17 12 42

AD C AD C

46 10 <5

AD I AD

71 200 30

C AD C

15 10–22

AD C

15–28

AD

Input material and disposal Collection materials Excluding emissions from collection/transports Excluding pretreatment (energy use) Excluding pretreatment (material losses) Treatment Substitution of coal power instead of coal lean power mix Substitution of coal power instead of coal lean power mix Use of coal power instead of coal lean power mix Excluding treatment of reject from pretreatment Excluding internal energy use Excluding internal energy use

Post-treatment Including CH4-emissions from storage of digestate Including direct N2O-emissions from storage of digestate Including carbon storage

Excluding emissions of N2O from farmland application (in substitution of fertilizers) Excluding NH3 emissions from bio-fertilizers Including indirect N2O emissions from organic fertilizers Excluding ash-treatment Excluding emissions from combustion of biogas Compensatory system Excluding substitution of chemical fertilizers/peat Including direct emissions of N2O from chemical fertilizers Including indirect emissions of N2O from chemical fertilizers 1 2

Assuming use of Swedish average power

Assuming 2% of organic waste ending up in bottom ashes and a storage factor of 25% (Smith et al. (2001), applied on landfills)

Following IPCC recommendations on indirect N2O-emissions2

Following IPCC recommendations on N2O-emissions.

Following Rodhe et al. (2009), Whitehead and Raistrick (1990), Sørensen and Birkmose (2002) and IPCC recommendations

Relative change in GWP in relation to results in the reference study related to inclusion/exclusion of a specific process for different treatment alternatives. Using IPCC recommendations on indirect N2O-emissions (0.01 kg N2O/kg emitted NH3) (IPCC, 2006).

when describing the importance of a parameter to the overall results from an LCA study, i.e. the robustness. In the present study the term large is used. Parameters where a change within the known uncertainty ratio or less than a 20% change in the specific parameter causes a change larger than 20% in relation to the overall outcome in relation to a specific environmental impact category (i.e. a sensitivity ratio > 20%) are described as large. The chosen method aims to test:

– the sensitivity of changes in relation to a specific parameter on the overall results from the study – the impact of changes in system boundaries, i.e. the effect of including/excluding specific parameters from the system studied in the reference paper Consequences are summarized in Tables 11 and 12 and discussed in relation to the outcomes from the literature review and guidelines in the following section.

2450

A. Bernstad, J. la Cour Jansen / Waste Management 32 (2012) 2439–2455

Table 12 Result from sensitivity analyses: relative change in input parameters needed to result in a 20% change in overall GWP from respective treatment alternative in the reference study. Input parameter Input material and disposal Increasing N-tot content in food waste Treatment Treatment of reject from pretreatment Internal energy use in incineration plant Total emissions of N Emissions of CH4

Emissions of N2O 1

Treatment alternative

Needed change (%)

C

25

AD

18

AD

50

I

90

C C

72 20

AD I C

80 >100 >100

Input parameter Compensatory system Emissions of N2O from production of chemical fertilizer Post treatment Substitution factor of N-fertilizers with organic fertilizers Direct emission of N2O from bio-fertilizers on farmland Increasing CH4 emissions from combustion of biogas

Treatment alternative

Needed change (%)

C

80

AD

35

C

–1

AD C

20 40

AD AD

18 >1000

A substitution ratio >100% would be needed to give a 20% change.

5. Discussion of key issues According to the ISO standard (ISO, 2000) an LCI should include relevant inputs and outputs of a product system and sensitivity analyses should focus on the most significant issues, to determine the influence on variations in assumptions, methods and data. However, defining what in- and outputs are to be seen as relevant and identifying the most significant issues is commonly not known in advance (Finnveden et al., 2009). In fact, sensitivity analyses have been proposed as a tool for identification of the significance of different issues (Eriksson and Baky, 2010; Björklund, 2002; Clavreul and Christensen, 2011), which shows the limited usefulness in the above cited ISO recommendation. The present paper show large differences in outcomes from previous LCAs of food waste management – both in absolute numbers and in ranking of alternatives. In line with Villanueva and Wenzel (2007) (reviewing a number of LCAs for paper waste management systems), noticed differences are not found to be due to the actual differences in the environmental impacts from studied systems, but rather to differences in system boundary setting and LCA methodology. Reasons for differences in results in the studies can be categorized as – Differences in system boundary setting – Methodological choices (for example assessing of GWP to biogenic carbon emissions or choice of substituted energy production) – Variations in used input data As previously stated by Morris (2011), including or excluding processes (referring to both emissions/resource use as well as benefits) can affect the estimated life cycle impacts and resultant ranking of one management method compared to another. Thus, the differences in relation to what is seen as significant/insignificant processes in the system boundary setting, using the terminology of Guinée et al. (2002), where parameters seen as highly relevant for the outcome of the study in some case not are considered at all in other studies, was seen as the major cause of differences in reviewed studies. Although system boundary setting always is closely connected to the aim of a specific study and it therefore is difficult to state any general indicators of how these should be set, system boundaries should always be handled with consistency over compared systems (Guinée et al., 2002). Koning et al. (2010) state that a fair comparison of two different products only could be made if the LCA background system is identical and exactly

the same choices are made in the foreground system. Since this seldom is the case, LCAs are unlikely to provide precise and meaningful information for comparisons between products, according to Koning et al. (2010). In order to increase the fairness in comparative LCAs or the comparability between different LCAs, standards with precise prescriptions on system boundary setting and choices must according to the same authors be generated. However, in comparison to product LCAs, LCAs of solid waste management systems are commonly more complex and several site dependent differences in background systems, such as energy production systems and climatic conditions can be of interest to capture in the study. Thus, an excessive streamlining is neither viable nor wanted. 5.1. Input material composition, collection/transportation and pretreatment A general difficulty when comparing LCAs of food waste management is that assumptions regarding the composition of the treated material could differ between different studies. It might be claimed that one should not compare studies with a varying composition of the examined waste fraction (in this case food waste). However, the fact that many researchers do not clearly define their input substrate composition is one of the observations of the present study. Several parameters, such as the carbon and nutrient content, but also assumed levels of toxic substances in treated waste have a high impact on GHG-emissions in relation to different treatment alternatives. Inspite of this, the present study shows that assumptions regarding these parameters in many cases are not displayed by LCA-study authors and vary largely amongst previously performed studies when included. Collection and transportation is of low relevance to the result, but was included in all reviewed LCAs with few exemptions (often with extensive calculations) and addressed in inquired guidelines. The influence of used collection material (including production and disposal of the same) was seen as negligible in the reference study, where paper bags where used. However, assuming use of plastic bags for food waste collection would give a 14–19% change of GWP in relation to AD and composting in the reference study (assuming combustion of both used plastic and paper bags with energy recovery). The choice of collection material could also affect the ratio of rejected material in pre-treatment or the usability of produced bio-fertilizers, which was not taken into consideration in sensitivity analyses made in the present study. Although not

2451

A. Bernstad, J. la Cour Jansen / Waste Management 32 (2012) 2439–2455

commonly addressed in LCAs of food waste, pretreatment can according to this study be of large importance. A material loss of 14% of input material (i.e. only 40–50% of the values presented by Truedsson (2010) based on a modern Swedish pretreatment facility) result in a 20% change in GWP from the AD alternative in the reference study.

Table 13 Emissions from storage of digestate; literature values and presentation of needed emission factors in order to reach a 20% change in relation to GWP in the reference study.

5.2. Treatment 1

Reduction of emissions of NH3, CH4 and N2O through the use of biofilters in composts has a large impact in relation to GWP both in reviewed studies and in relation to the reference study. Reducing the emissions of NH3 and CH4 with 95% and 25% respectively (possible with biofilters according to Liang et al. (2000), Busca and Pistarino (2003), Dalemo et al. (1997)) would result in a 25% change in relation to GWP. In line with Lee et al. (2007) it was seen that the impact from the use of structural material was insignificant in relation to GWP. The influence associated to energy input in treatment processes will to a large extent depend on assumptions related to the eco-profile of used energy. Due to the low calorific value in wet food waste several authors argue that food waste cannot be incinerated independently from other waste (Smith et al., 2001; Chaya and Gheewala, 2007; Diggelman and Ham, 2003). If support fuel is used, system boundary extension is needed in order to address the environmental impacts connected to production of this fuel and emissions from combustion of the same should be addressed. Food waste is normally co-combusted with other waste fractions, with higher energetic values. It could therefore be stated that the comparison between thermal treatment and other treatment alternatives refers to combustion of an extra specific amount of food waste combusted, rather than combustion of food waste alone. If this extra amount is relatively small, it will probably not affect the overall efficiency of the incineration plant. However, with an increasing amount of food waste in the waste composition, key factors such as the electricity recovery factor could be affected. Thus, a support fuel is always needed, and if this is assumed to be mixed MSW, it could be argued that this decreases incentives for material recycling. Several previous studies have shown that material recycling commonly is superior to incineration in relation to environmental impacts (Villanueva and Wenzel, 2007; Ekvall et al., 1999). Thus, it could be argued that incineration of food waste should be burdened with environmental impacts from a nonachieved recycling of waste such as plastics and paper (i.e. materials which could be recycled are not, as they are needed to increase the heating value in combusted waste). As argued by Morris (2010), these could be viewed as upstream burdens from using potentially recyclable materials for energy production rather than recycling when needed for combustion of food waste. 5.3. Post treatment and use of produced goods Emissions from storage of biotreated materials were not included in the reference study. A methane conversion factor (i.e. production of methane from potentially anaerobically degradable organic matter) of 10% is needed in order to give a 20% change in GWP in relation to the reference study, while direct and indirect emissions of N2O from storage of bio-fertilizers would have to be much higher than assumptions made in previous studies in order to be relevant in relation to GWP results (Table 13, based on an assumed degradation ratio of biodegradable material of 75–85%. Post-treatment processes of secondary waste streams are commonly not included in reviewed studies when food waste is incinerated (Table 9). Thus, there is a potential risk of biased comparisons if post-treatment of ashes is excluded while use of secondary waste (bio-fertilizers) from biological treatment alternatives is considered. This is relevant not only in cases were

2 3

Emission

Literature values

Needed value (%)

Unit

CH4 N2O N2O

15%1, 2.7%2 0.1%1, 0.5%3 1%2

10 1.7 19

% Of remaining CH4 % Of N-tot % Of NH3 emissions

direct indirect

Lantz et al. (2009). IPCC (2006). Kärrman et al. (2005).

Fig. 3. GWP from biological treatment alternatives in studies as total average and average with different limitations.

toxicity is addressed as an impact category, as the present study show that use of bio-fertilizers as substitution for chemical fertilizers and soil improvement media can have a large impact also on other impact categories, such as GWP. Whether benefits from substitution of chemical fertilizers from biological treatment alternatives was taken into consideration in the study or not has a clear impact on the GWP, while the impact from carbon sequestration seems to be of lesser importance (Fig. 3). Assuming that produced compost and digestate are not suitable as fertilizer is often related to and assumed high amount of toxic compounds in treated waste. Comparing assumptions used in reviewed studies and consulted guidelines it is clear that these assumptions vary largely (Pb 3.9–31.6; Cd 0.10–0.41; Hg 0.02– 0.12; Cu 16–52; Cr 7–16 and Zn 45–197 mg/kg TS in Hirai et al., 2000; Kärrman et al., 2005; Bjarnadottir et al. 2002; Fruergaard and Astrup, 2011). Independent of the high uncertainties related to the metal content in food waste, assessing the impact from different assumptions are rare (Kirkeby et al., 2006). Emissions from soil application of produced digestate and compost are dependent on several factors, such as spreading technique (Stangel, 1988; Nanh et al., 2008), application procedure and climate/season (Dalemo et al., 1997), the total nutrient content and the nitrogen and carbon content matrix, soil conditions, degradation of organic pollutions and the transfer coefficients of metals in soil etc. Emission factors for direct N2O emissions from bio-fertilizers on farmland used in the reference study are based on Bouwman et al. (2002), presenting a span between 0.02% and 2.5% of total N-content in fertilizers. Applying the lower value or assuming that no N2O is emitted (the case in all but three reviewed studies) increases the overall avoidance of GWP with 30 and 60% in relation to the composting and AD scenarios described in the reference study respectively. In the reference study, emissions of NO3 from bio-fertilizers were assumed to 20% and volatilization of NH3 to 7.5% of N-tot based on Bruun et al. (2006). None of the reviewed studies take indirect emissions of N2O from utilization of bio-fertilizers into consideration. Including this in the reference study decreases net GWP avoidance with 34 and 36% in relation to AD and composting respectively.

2452

A. Bernstad, J. la Cour Jansen / Waste Management 32 (2012) 2439–2455

Emissions of NH3, NO3 or N2O (direct as well as indirect) from application of chemical fertilizers on farmland were not included in the reference study, nor addressed in any of the reviewed studies. Volatilization of NH3 from chemical fertilizers has previously been assumed to 0, 1 and 53% of applied N-tot (Bruun et al., 2006; Rodhe et al., 2009; Whitehead and Raistrick, 1990 respectively). The amount of NO3 leached from chemical fertilizers has been estimated to 10% lower compared to use of digestate (Sørensen and Birkmose, 2002). Including indirect emissions of N2O related to volatilization of NH3 based on the data above result in changes in GWP between 10–22% and 15–28% in relation to AD and composting respectively. As for direct emissions of N2O, IPCC recommendations state that the same emission factors can be applied for all types of fertilizers (i.e. also in relation to substituted chemical fertilizers), which would result in an increase in reduction of GWP with 15–30% in relation to compost and AD respectively. 5.4. Compensatory system Changes in the compensatory systems can result in large changes in the overall results in the reference study. While the importance of the eco-profile of substituted electrical energy is much discussed and commonly investigated through sensitivity analyses, the same is not seen in relation to other goods substituted by the different waste treatment options, such as thermal energy and chemical fertilizers. Production of chemical fertilizers has a large impact on GWP, mainly due to the energy consumption for production of ammonia, which to 99% is covered using fossil energy, and emissions of N2O (IFA, 2011). However, several abatement technologies for N2O emissions are currently on the market (primary, secondary, non-selective catalytic reactions (NSCR) and tertiary control measures), which can reduce N2O emissions with up to 90% (EPA, 2010). Datasets for production of substituted chemical fertilizers are often outdated (1996 in the case of Bovea and Powell (2006) and 1999 in the case of Börjesson and Berglund (2007)). Assuming that the substituted fertilizers are produced with BAT (based on Jensen and Kongshaug (2003), see section 3.5. Substituted goods), reduces avoided GWP with 54% and 12% in relation to AD and composting respectively in the reference study. 5.5. Performance of sensitivity analyses Sensitivity analyses in reviewed studies most commonly addressed the eco-profile of substituted electrical energy, treatment emissions, energy recovery rate, transports and potential carbon sequestration. Sensitivity analyses in the same studies as well as in the reference study show that transports and carbon sequestration commonly are of little importance to the results. Changes in the food waste composition are very seldom assessed and no cases were found where emissions from farmland application were addressed through sensitivity analyses, although the reference study shows that such changes can have a large effect. 6. Methodological discussion The combination of analysing the needed parameter change in input data in order to create a 20% impact on the overall results together with the analyse of the effect of exclusion of a specific parameter was seen as useful in key issue identification as this method allows both for testing the influence of system boundary setting choices as well as the importance of large variability in used input data. However, in some cases, such as fugitive emissions of methane from AD plants, the needed parameter change

to result in a 20% change in relation to GWP is large (80%) and exclusion of this emission does not lead to a large change on the overall GWP from the AD system. This could be interpreted as if fugitive emissions from AD plants are irrelevant and that such could be excluded from the systems analysis without any larger impact. However, the results should be seen in relation to the level of fugitive emission assumed in the base case (1% of produced CH4). Had initial assumptions stated higher emissions, would needed changes in relation to this parameter in order to reach a 20% change in overall outcomes have been lower. With the 1% emission assumed in the base case, the resulting emission factor (when increased by 80%) is still within what previously has been seen as reasonable according to IPCC (2006) and Swedish Waste Management Association (2007). Thus, it could be of relevance to assess the influence of even higher emissions and a third method for key parameter testing could be a combination of applying maximum and minimum values (based on literature data) with inclusion/exclusion of a specific process in the system. Changes in outcomes depending on changes in input parameters is in the present study in all cases reported as relative values, i.e. as percent of the initial output value. Thus, the observed importance of parameter-changes will depend on the initial output value. For example, in a system where transportation is inefficient and the impact from this part of the treatment chain is large, the relative importance of other parameters, such as indirect emissions of N2O from digestate on farmland will decrease, while in a system where emissions from bio-fertilizers on farmland was not taken into consideration, the relative importance of emissions from transportations are likely to increase. Also, other processes in some cases counteract changes in the system leading to increased impacts, such as in the case of NH3 volatilization from storage of digestate, which decreases later emissions from farmland. This drawback has earlier been identified by Björklund (2002), stating that potential synergisms are overlooked when using sensitivity analyses for key issue identification and estimation of uncertainties. As an example, internal energy use in treatment facilities can be seen both as having a high importance to the overall GWP from a specific scenario and not, depending on the eco-profile of the used energy. 6.1. Suggestions and further research A general conclusion based on the performed review is that differences in results seen in compared studies to a large extent can be explained by differences in relation to used system-boundary settings and assumptions made in respective study, rather than actual differences in relation to compared treatment alternatives. Results from the present study are thereby in line with previous findings in relation to MSW LCA-studies (Gentil et al., 2009). Thus, results do not state that it is not advisable to use LCA as a decision support tool for food waste management, due to the fact that results in reviewed studies vary largely, but rather that there is large room for improvements and methodological streamlining in order to improve the general quality as well as potentials for cross-study comparisons. This review includes studies performed in relation to waste management systems in different countries, in many cases with very different levels of technological development and climate conditions. As many biological processes are strongly climate dependent (mainly in relation to temperature and humidity), it could be expected that results from different geographical areas would be influenced by such factors. The same could be said in relation to technical developments, influencing efficiency in energy/nutrient recovery, emissions from machinery etc. However, the review shows that such geographical considerations seldom are taken into consideration, as used LCI-data commonly consist

A. Bernstad, J. la Cour Jansen / Waste Management 32 (2012) 2439–2455

of literature data from other regions or non-country specified data collected from LCA databases. The risk with such practices is highlighted in ILCD (2010b) and this problem is not unique in the area of food waste management LCAs. However, due to the biologically degradable nature of food waste in relation to other waste fractions, such as plastics or glass, the importance of taking such factors into consideration could be even more important within this area. The different result presented in reviewed studies can to a large extent be explained by differences in the view on what is actually compared, i.e. treatment of the functional unit (FU) in a business as usual scenario (typically landfill or incineration of MSW) or treatment of 1 FU separately. As an example, in the first case, combustion is accomplished together with other waste, without need for support fuel, allowing crediting of energy recovery. In the second case, combustion with energy recovery is only accomplished if support fuel is added, which will change the environmental profile of the same treatment alternative drastically. Also, in the second case, all transports and material-use related to selective collection will burden treatment alternatives related to selective collection, while co-collection of food waste with other waste typically is not burdening the food waste treatment. Although food waste never would be neither combusted nor landfilled alone in practice, this paper argues that the latter reflects a more un-biased basis for comparison. Yet another way to solve this dilemma would be to include environmental burdens related to the management chain for non-food waste needed in co-treated of food waste to make the suggested treatment alternative practically viable. Hence, based on this review, the following suggestions can be made: – Increase the knowledge of the characteristics of the treated material; primarily carbon content and matrix, nutrient content and content of toxic elements. Emissions of nitrogen and carbon containing compounds, potentials for soil application of bio-fertilizers, characteristics in leachate and ashes, potential energy recovery and potential carbon sequestration will to a large extend depend on the characteristics of treated material. – Apply mass balance calculations for substances (primarily for carbon, nutrients and heavy metals) in the system and make use of transfer coefficients rather than literature references in order to maintain mass balances over the treatment process. – Improve coherence in assigning of GWP to biogenic carbon emissions. Recommendations on how this can be done are found in Christensen et al. (2009). – Assure internal consistency in system boundary setting and use of cut-offs between compared alternatives in order to assure unbiased comparisons. Emissions that would have emerged also in a reference scenario should be excluded from non-reference scenarios, alternatively included in both cases. – Increase the use of sensitivity analyses in order to assess the impact of including/excluding different processes as well as use of differing input values in relation to processes were uncertainties are large. – Reflect on the need to adjust used input data in relation to geographical and technical conditions in the specific case. The study also point out a number of areas which, according to performed sensitivity analyses, can be of high relevance to the overall results of the LCA and where further research therefore is highly motivated in order to improve input data quality and robustness of results: – Potential emissions of carbon, nutrients and others in food waste during pre-collection, collection and transportation – Losses during physical pretreatment of food waste (rejects)

2453

– Potentials for energy recovery through incineration of food waste – Ash generation (quantity and characteristics) from food waste incineration – Direct emissions from composting of food waste under different circumstances – Emissions from bio-fertilizers and chemical fertilizers during storage and land application under different circumstances As stated above, none of the reviewed studies have a dynamic perspective, giving that the effects of changing between different treatment-options in a larger perspective are not recognized. As an example; lifting 1 ton food waste from the current waste treatment chain (often landfill or incineration) to biological alternatives will have an impact on the heating value (amongst others) in the waste going to incineration and processes in the landfill, which commonly is not reflected. LCA methodology is currently unable to address many impacts related to land application of bio-fertilizers, such as improved soil quality and increased water retention capacity. This could be related to the current lack of methods to address such impacts in numeric terms. Although example exists of previous LCA studies, trying to address the impact from increased water retention through application of bio-solids on farmland, the uncertainties regarding long term effects, plant systems and local conditions are large, resulting in the need for many assumptions (Peters and Rowley, 2009). Two guidelines suggest mentioning of this type of impacts qualitatively in the LCA (Bjarnadottir et al., 2002; ILCD, 2011a (in relation to compost but not to digestate)). Resource use is seldom addressed in the studies. Currently available characterization models used in LCIA for resources often have an anthropocentric approach as they focus on the use value for humans and exclude their intrinsic value and the non-use of the resource. Phosphorus is probably one of the most relevant resources related to organic waste treatment and is of relevance both from an anthropocentric viewpoint and through the potential benefits related to conservation of natural environment connected to non-use of mineral stocks. Another important issue is the close relationship between phosphorus mining and increased spreading of cadmium in the biosphere (Packman, 2011). However, nutrient recycling made possible through biological treatment methods is in reviewed studies credited only as energy savings and reduced emissions from production of substituted of chemical fertilizers. A challenge for the future is therefore development of quantitative methods for inclusion also of benefits related to other perspectives of an increased fertilizer recovery. Another future challenge is addressing the site-specific nature of agricultural systems. These systems are often an important part of the LCA of organic waste management if the use of produced organic fertilizers is taken into consideration. This encourages an increased regionalization of emissions and characterization factors for EP and AP caused by release of N and P from agricultural systems, as previously stated by Hayashi et al. (2005).

7. Conclusions This review of twenty-five life cycle assessments of different food waste treatment alternatives shows that both absolute values and relative ranking of compared treatment alternatives differs largely in relation to their impact on global warming potential. However, these differences are not found to be related mainly to actual differences in the environmental impacts from studied systems, but rather to differences in system boundary settings, methodological choices and, to a lesser extent, to variations in used input data. Due to lack of consistency in system boundary setting, processes

2454

A. Bernstad, J. la Cour Jansen / Waste Management 32 (2012) 2439–2455

and parameters seen as highly relevant for the outcome in some studies are not considered at all in others, making comparisons between different studies difficult. Differences in system boundary setting were also seen internally, between compared treatment alternatives in the same study, resulting in internal inconsistencies and biased comparisons. Assumptions related to the characteristics of treated food waste, losses of carbon, nutrients and other compounds during physical pretreatment, potential energy recovery through incineration, emissions from composting, emissions from storage and land use of bio-fertilizers and chemical fertilizers and eco-profiles of substituted goods were all identified as highly relevant for the outcomes in relation to global warming from this type of comparative life cycle assessments. Mass-flows of carbon, nutrients and heavy metals are in many cases not respected due to cut-offs and use of literature values rather than transfer coefficients throughout the treatment chain. Physical pretreatment of food waste, storage of bio-fertilizers, emissions from soil application of chemical fertilizers, use of support fuel and treatment of incineration residues were commonly not addressed in reviewed studies, although sensitivity analyses performed in the present work show that they can be of large importance. These are also in many cases areas where there is a need for further research and improvement of input data. The six life cycle assessment guidelines consulted in this work did in many cases not address several of the identified key issues or give different recommendations in relation to the same. Thus, establishment and use of more detailed guidelines within this field in order to increase both the general quality in assessments as well as increase the possibilities to compare different studies would be welcomed. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.wasman.2012. 07.023. References Aye, L., Widjaya, E.R., 2005. Environmental and economic analyses of waste disposal options for traditional markets in Indonesia. Waste Management 26, 1180– 1191. Baky, A., Eriksson, O., 2003. Systems analysis of organic waste management in Denmark. Environmental Project No. 822. Danish Environmental Protection Agency, Copenhagen, Denmark. Banar, M., Cokaygil, Z., Ozkan, A., 2009. Life cycle assessment of solid waste management options for Eskisehir, Turkey. Waste Management 29, 54–62. Benjaminsson, J., 2006. Nya uppgraderingstekniker för biogas/New upgrading techniques for biogas. Master Thesis. LITH-IKP-EX-06/2370—SE. Department of Mechanical Engineering, Lindköping University, Sweden (in Swedish). Bernstad, A., la Cour Jansen, J., 2011. A life cycle approach to the management of household food waste – a Swedish full-scale case study. Waste Management 31, 1879–1896. Bjarnadottir, H.J., Fridriksson, G.B., Johnsen, T., Sletnes, H., 2002. Guidelines for the Use of LCA in the Waste Management Sector. Nordtest Report TR 517. Nordtest, Espoo, Finland. Björklund, A., 2002. Survey of approaches to improve reliability in LCA. International Journal of LCA 7 (2), 64–72. Blengini, G.A., 2008a. Applying LCA to organic waste management in Piedmont, Italy. Organic waste management 19 (5), 533–549. Blengini, G.A., 2008b. Using LCA to evaluate impacts and resources conservation potential of composting: a case study of the Asti District in Italy. Resources, Conservation and Recycling. 53, 1373–1381. Boldrin, A., Andersen, J.K., Møller, J., Christensen, T.H., Favoino, E., 2009. Composting and compost utilization: accounting of greenhouse gases and global warming contributions. Waste Management and Research 27, 800–812. Börjesson, P., Berglund, M., 2007. Environmental systems analysis of biogas systems—Part II: The environmental impact of replacing various reference systems. Biomass and Bioenergy 31, 326–344. Bouwman, A.F., Boumans, L.J.M., Batjes, N.H., 2002. Modelling global annual N2O and NO emissions from fertilised fields: summary of available measurement data. Global Biogeochemical Cycles 16 (4), 28–31. Bovea, M.D., Powell, J.C., 2006. Alternative scenarios to meet the demands of sustainable waste management. Journal of Environmental Management 2, 115– 132.

Bruun, S., Lund Hansen, T., Christensen, T.H., Magid, J., Jensen, L.S., 2006. Application of processed organic municipal solid waste on agricultural land – a scenario analysis. Environmental Modelling and Assessment 11 (3), 251–265. Busca, G., Pistarino, C., 2003. Abatement of ammonia and amines from waste gases: a summary. Journal of Loss Prevention in the Process Industries 16, 157–163. Cabaraban, M.T.I., Khire, M.V., Alocilja, E.C., 2008. Aerobic in-vessel composting versus bioreactor landfilling using life cycle inventory models. Clean Technologies and Environmental Policy 10 (1), 39–52. Chaya, W., Gheewala, S.H., 2007. Life cycle assessment of MSW-to-energy schemes in Thailand. Journal of Cleaner Production 15, 1463–1468. Cherubini, F., Bargigli, S., Ulgiati, S., 2009. Life cycle assessment (LCA) of waste management strategies: landfilling, sorting plant and incineration. Energy 34 (12), 2116–2123. Christensen, T.H., Emmanuel, G., Boldrin, A., Larsen, A.W., 2009. C balance, carbon dioxide emissions and global warming potentials in LCA-modelling of waste management systems. Waste Management and Research 27 (8), 707–715. Clavreul, J., Christensen, T.H., 2011. Uncertainties in life cycle assessment of waste management systems. In: Proceedings from the Thirteenth International Waste Management and Landfill Symposium, 2011, Sardinia, Italy. Clemens, J., Trimborn, M., Weiland, P., Amon, B., 2006. Mitigation of greenhouse emissions by anaerobic digestion of cattle slurry. Agriculture Ecosystems & Environment 112, 171–177. Clift, R., Doig, A., Finnveden, G., 2000. The application of life cycle assessment to integrated solid waste management. Process safety and environmental protection. Transactions of the Institution of Chemical Engineers 78 (4), 279– 287. Dalemo, M., Sonesson, U., Bjorklund, A., Mingarini, K., Frostell, B., Jonsson, H., Nybrant, T., Sundqvist, J.-O., Thyselius, L., 1997. ORWARE – a simulation model for organic waste handling systems. Part 1: Model description. Resources, Conservation and Recycling 21, 17–37. Diggelman, C., Ham, R., 2003. Household food waste to wastewater or to solid waste? That is the question. Waste Management and Research 21 (6), 501– 514. Eggleston, S., Buendia, L., Miwa, K., Ngara, T., Tanabe, K., 2006. IPCC Guidelines for National Greenhouse Gas Inventories. Vol. 5 Waste. IPCC National Greenhouse Gas Inventories Programme. Institute for Global Environmental Strategies, Hayama, Kanagawa, Japan. EPA, 2010. Available and emerging technologies for reducing greenhouse gas emissions from the nitric acid production industry. Sector Policies and Programs Division Office of Air Quality Planning and Standards US Environmental Protection Agency Research Triangle Park, North Carolina 27711, US. Eriksson, O., Baky, A., 2010. Identification and testing of potential key parameters in system analysis of Municipal solid waste management. Conservation and Recycling 4, 1095–1099. European Parliament, 2008. Directive 2008/98/EC of the European Parliament and the of the council of 19 November 2008 on waste and repealing certain Directives. Finnveden, G., 1992. Landfilling - a forgotten part of Life Cycle Assessments. Product Life Cycle Assessments - Principles and Methodology, vol. 9. Nordic Council of Ministers, Copenhagen, Denmark, pp. 263–280. Finnveden, G., Hauschild, M.Z., Ekvall, T., Guinee, J., Heijungs, R., Hellweg, S., Koehler, A., Pennington, D., Suh, S., 2009. Recent developments in Life Cycle Assessment. Journal of Environmental Management 91 (1), 1–21. Fruergaard, T., Astrup, T., 2011. Optimal utilization of waste-to-energy in an LCA perspective. Waste Management 31 (3), 572–582. Gentil, E.C., Damgaard, A., Hauschild, M., Finnveden, G., 2010. Models for waste life cycle assessment: Review of technical assumptions. In: Eriksson, O., Thorneloe, S., Kaplan, P.O., Barlaz, M., Muller, O., Matsui, Y., Ii, R., Christensen, T.H. (Eds.), Waste Management 30 (12), 2636–2648. Güereca, L.P., Gassó, S., Baldasano, J.M., Jiménez-Guerrero, P., 2006. Life cycle assessment of two biowaste management systems for Barcelona, Spain. Resources, Conservation and Recycling 49, 32–48. Guinée, J.B. et al., 2002. In: Guinée, J.B. (Ed.), Handbook on Life Cycle Assessment Operational Guide to the ISO Standards. Kluwer, Dordrecht, Boston, London. Hansen, T., la Cour Jansen, J., Davidsson, Å., Christensen, T.H., 2007. Effects of pretreatment technologies on quantity and quality of source-sorted municipal organic waste for biogas recovery. Waste Management 27 (3), 398–406. Hayashi, K., Gaillard, G., Nemecek, T., 2005. Life cycle assessment of agricultural production systems: current issues and future perspectives. In: International Seminar on Technology Development for Good Agriculture Practice in Asia and Oceania, 25–26th of October, 2005, Tsukuba, Japan. Heijungs, R., Huijbregts, A.J., 2004. A review of approaches to treat uncertainties in LCA. In: Proceedings of the 2nd Biennial Meeting of iEMSs, Complexity and integrated resources management, Osnabrück, Germany, 14–17 June, 2004, pp. 332–339. Hirai, Y., Murata. M., Sakai, S., Takatsuki, H., 2000. Life Cycle Assessment for Food Waste Recycling and Management. In: Proceedings from the 4th International Conference on EcoBalance Methodologies for decision Making in a Sustainable 21st Century. October 31st–November 2nd, 2000, Tsukuba, Japan. IFA, 2011. International Fertlizer Industry Association. Climate Change. (Downloaded 20110925). ILCD, 2010. ILCD handbook. International Reference Life Cycle Data System. General guide for Life Cycle Assessment – Detailed guidance. EUR 24708 EN – 2010. JRC, European Commission.

A. Bernstad, J. la Cour Jansen / Waste Management 32 (2012) 2439–2455 ILCD, 2011a. Support JRC Scientific and Technical Reports. Supporting Environmentally Sound Decisions for Waste Management – A Technical Guide to Life Cycle Thinking (LCT) and Life Cycle Assessment (LCA) for Waste Experts and LCA Practitioners. ILCD, 2011b. Support JRC Scientific and Technical Reports. Supporting Environmentally Sound Decisions for Food Waste Management – A Technical Guide to Life Cycle Thinking (LCT) and Life Cycle Assessment (LCA) for Waste Experts and LCA practitioners. IPCC, 2006. IPCC 2006 Guidelines for national greenhouse gas inventories. In: Eggleston, H.S., Buendia, L. Miwa, K. Ngara, T. och Tanabe, K. (Eds), Prepared by the National Greenhouse Gas Inventories Programme, vol. 4. Agricultural, Forestry and Other Land Use, IGES, Japan. ISO, 2000. International Organization for Standardization. Environmental management – Life cycle assessment - Life cycle impact assessment. ISO 14042:2000. (accessed 15.10.11). ISO, 2006a. International Organization for Standardization. Environmental management – Life cycle assessment – Principles and framework, second ed., ISO 14040:2006. (accessed 15.10.11). ISO, 2006b. International Organization for Standardization. Environmental management – Life cycle assessment – Requirements and guidelines, first ed., ISO 14044:2006. (accessed 15.10.11). Jensen, T.K., Kongshaug, G., 2003. Energy consumption and greenhouse gas emissions in fertiliser production. In: Proceedings No. 509, Paper presented to The International Fertiliser Society at a Meeting in London, on 3rd of April 2003. The International Fertilizer Society, York, United Kingdom. ISBN 0 85310 145 0. Kirchmann, H., 1985. Household Composts, Waste-Water Treatment Sludge and Byproducts from Industry in Agriculture; Problems, Demand and Research. ISBN 91-576-3283–9. SLU Uppsala, Sweden (in Swedish). Khoo, H.H., Lim, T.Z., Tan, R.B.H., 2010. Food waste conversion options in Singapore: environmental impacts based on an LCA perspective. Science of the Total Environment 408, 1367–1373. Kim, M.-H., Kim, J.-W., 2010. Comparison through a LCA evaluation analysis of food waste disposal options from the perspective of global warming and resource recovery. Science of the Total Environment 408, 3998–4006. Kirkeby, J.T., Birgisdottir, H., Lund Hansen, T., Christensen, T.H., Bhander, G.S., Hauschild, M., 2006. Evaluation of environmental impacts from municipal solid waste management in the municipality of Aarhus, Denmark (EASEWASTE). Waste Management and Research 24, 16–26. Koning, A., Schowanek, D., Dewaele, J., Weisbrod, A., 2010. Uncertainties in a carbon footprint model for detergents; quantifying the confidence in a comparative result. The International Journal of Life Cycle Assessment 15 (1), 79–89. Kärrman, E., Baky, A., Edström, M., Magnusson, Y., Malmqvist, P.-A., Palm, O., Rogstrant, G., 2005. Systemstudie rörande insamling och behandling av lättnedbrytbart organiskt avfall i Malmö/Systems study on collection and treatment of easily degradable orgnaic wast in Malmö. Report 2005004, Ecoloop, 2005, Stockholm, Sweden (in Swedish). la Cour Jansen, J., Christensen, T., Davidsson, Å., Lund Hansen, T., Jönsson, H., Kirkeby, J., 2007. Biowaste – Decision Support Tool for Collection and Treatment of Source-Sorted Organic Municipal Solid Waste. TemaNord 2007:602. Nordic Council of Ministers, Copenhagen. Lantz, M., Ekman, A., Börjesson, P., 2009. Systemoptimerad produktion av fordonsgas – En miljö- och energioptimerad studie av Söderåsens biogasanläggnin/Systems optimized production of vehicle gas – an environmental and energy assessment of the Söderåsen biogas production plant. Report 69. Envionmental and Energy Systems Studies, Lund University, Lund Sweden (in Swedish). Lee, S.-H., Choi, K.-I., Osako, M., Dong, J.-I., 2007. Evaluation of environmental burdens caused by changes of food waste management systems in Seoul, Korea. Science of the Total Environment 387, 42–53. Liang, Y., Quan, X., Chen, J., Chung, J.S., Sung, J.Y., Chen, S., Xue, D., Zhao, Y., 2000. Long-term results of ammonia removal and transformation by biofiltration. Journal of Hazardous Materials 80, 259–269. Linzner, R., Mostbauer, P., 2005. Composting and its impact on climate change with regard to process engineering and compost application – a case study in Vienna. In: Proceedings of Sardinia 2005 – Tenth International landfill Symposium. CISA Publishing, Cagliari. Martin, J.P., Haider, K., 1971. Microbial activity in relationship to soil humus formation. Soil Science 111, 54–63. Martínez-Blanco, J., Colón, J., Gabarrell, X., Font, X., Sánchez, A., Artola, A., Rieradevall, J., 2010. The use of life cycle assessment for the comparison of biowaste composting at home and full scale. Waste Management 30, 983–994. Mathiesen, B.V., Münster, M., Fruergaard, T., 2009. Uncertainties related to the identification of the marginal energy technology in consequential life cycle assessments. Journal of Cleaner Production 17, 1331–1338. Møller, J., Boldrin, A., Christensen, T.H., 2009. Anaerobic digestion and digestate use: accounting of greenhouse gases and global warming contribution. Waste Management and Research 27, 813–824. Morris, J., 2010. Bury or burn North America MSW? LCAs provide answers for climate impacts & carbon neutral power potential. Environmental Science and Technology 44, 7944–7949. Morris, J., 2011. Review of LCAs on Organics Management Methods and Development of an Environmental Hierarchy. Alberta Environment, Edmonton, AB.

2455

Nanh, C., Warren, J., Sistani, K., 2008. Ammonia and Greenhouse Gases Emission from Land Application of Swine Slurry: A Comparison of Three Application Methods. American Society of Agricultural and Biological Engineers, St. Joseph, Michigan. . Nielsen, O.-K., Lyck, E., Mikkelsen, M.H., Hoffmann, L., Gyldenkærne, S., Winther, M., Nielsen, M., Fauser, P., Thomsen, M., Plejdrup, M.S., Illerup, J.B., Sørensen, P.B., Vesterdal, L., 2008. Denmark’s National Inventory Report 2008 - Emission Inventories 1990–2006. Submitted under the United Nations Framework Convention on Climate Change. NERI Technical Report No. 667. National Environmental Research Institute, University of Aarhus, Denmark. . Packman, H., 2011. Risk Assessment of Cadmium in the Swedish Environment. Swedish Chemicals Agency, Sundbyberg, Sweden. Pertl, A., Mostbauer, P., Obersteiner, G., 2010. Climate balance of biogas upgrading systems. Waste Management 30, 92–99. Peters, G.M., Rowley, H.V., 2009. Environmental comparison of biosolids management systems using life cycle assessment. Environmental Science and Technology 43 (8), 2674–2679. PCR, 2008. Product Category Rules (PCR) for preparing an Environmental Product Declaration (EPD) for Solid Waste Disposal Services. PCR 2008:02. . Rigamonti, L., Grosso, M., Giugliano, M., 2009. Life cycle assessment for optimising the level of separated collection in integrated MSW management systems. Waste Management. 29, 934–944. Rodhe, L., Ascue, J., Tersmeden, M., Ringmar, A., Nordberg, Å., 2009. Växthusgaser från lager med nötflytgödsel – Förhållanden i gårdsbehållare, metodikutveckling av gasmätning samt bestämning av emissioner från nötflytgödsel, /Greenhouse gas emissions from storage of mannure – Conditions in farm storage, method development of gas emission detection and determination of emissions from liquid mannure. Report No. 370JTI, Uppsala, Sweden (in Swedish). Saltelli, A., Andres, T.H., Homma, T., 1995. Sensitivity analysis of model output. Performance of the iterated fractional factorial design method. Computational Statistics and Data Analysis 20 (4), 387–407. Smith, A., Brown, K., Ogilvie, S., Rushton, K., Bates, J., 2001. Waste Management Options and Climate Change. Final report to the European Commission, DG Environment. Office for Official Publications of the European Communities, Luxembourg. Sonesson, U., Björklund, A., Carlsson, M., Dalemo, M., 2000. Environmental and economic analysis of management systems for biodegradable waste. Resources, Conservation and Recycling 28, 29–53. Stangel, P., 1988. Technological options affecting emissions. Paper Presented at US Environmental Protection Agency Workshop on Agriculture and Climate Change, February 28–March 1, 1988, Washington DC. Sundqvist, J.-O., 1999. Life cycles assessments and solid waste – Guidelines for solid waste treatment and disposal in LCA. IVL, Swedish Environmental Research Institute. Report 279. AFN, Swedish Environmental Protection Agency, Stockholm, Sweden. Swedish Waste Management Association, 2007. Frivilligt åtagande – inventering av utsläpp från biogas- och uppgraderingsanläggningar/Mapping of methane emissions from Swedish biogas plants (in Swedish). Report U2007:02. Malmö, Sweden. Swedish Waste Management Association, 2010. Viktreducering, energiförlust och gasemissioner vid olika insamlingssystem av matavfall från hushåll/Weight reduction, energy losses and gas emissions in different systems for collection of household food waste (in Swedish). Report U2010:10. ISSN 1103–4092. Malmö, Sweden. Sørensen, P., Birkmose, T., 2002. Kvaelstofudvaskning efter gødskning med afgasset gylle /Nitrogen leachate after fertilizing with digestate from cow manure. Grøn Viden, Markbruk nr 266, Ministeriet for Fødevarer, Landbrug og Fiskeri, Danmarks Jordbruksforskning, Copenahgen, Denmark (in Danish). Tan, R.B.H., Khoo, H.H., 2006. Impact assessment of waste management options in Singapore. Journal of Air and Waste Management Association 56, 244–254. Truedsson, C., 2010. Positiva och negative miljöaspekter genom optimering av förbehandling av matavfall från hushåll/Positive and Negative Environmental Aspects through Optimization of Food Waste Pretreatment at Sysav Biotec. Master Thesis. Department for Chemical Engineering, Lund University, Sweden (in Swedish). Tyskeng, S., Finnveden, G., 2007. Energy and environmental differences between material recycling and energy utilization of waste – a literature review. TRITAINFRA-FMS 2007:11, KTH, Stockholm (in Swedish). USEPA, 1995. Guidelines for Assessing the Quality of Life-Cycle Inventory Analysis. Office of Solid Waste. US Environmental Protection Agency, Washington, DC, USA. Villanueva, A., Wenzel, H., 2007. Paper waste – recycling, incineration or landfilling? A review of existing life cycle assessments. Waste Management 27 (8), S29–S46. Whitehead, C.C., Raistrick, N., 1990. Ammonia volatilization from five nitrogen compounds used as fertilizers following surface application to soils. European Journal of Soil Science 41 (3), 380–394. Winkler, J., Bilitewski, B., 2007. Comparative evaluation of life cycle assessment models for solid waste management. Waste management 27 (8), 1021–1031.