Food waste minimization from a life-cycle perspective

Journal of Environmental Management 147 (2015) 219e226

Contents lists available at ScienceDirect

Journal of Environmental Management journal homepage: www.elsevier.com/locate/jenvman

Food waste minimization from a life-cycle perspective A. Bernstad Saraiva Schott*, T. Andersson Water and Environmental Engineering, Lund University, Kemicentrum, Box 124, 210 00 Lund, Sweden

a r t i c l e i n f o

a b s t r a c t

Article history: Received 21 February 2013 Received in revised form 18 July 2014 Accepted 22 July 2014 Available online 26 September 2014

This article investigates potentials and environmental impacts related to household food waste minimization, based on a case study in Southern Sweden. In the study, the amount of avoidable and unavoidable food waste currently being disposed of by households was assessed through waste composition analyses and the different types of avoidable food waste were classified. Currently, both avoidable and unavoidable food waste is either incinerated or treated through anaerobic digestion. A hypothetical scenario with no generation of avoidable food waste and either anaerobic digestion or incineration of unavoidable food waste was compared to the current situation using the life-cycle assessment method, limited to analysis of global warming potential (GWP). The results from the waste composition analyses indicate that an average of 35% of household food waste is avoidable. Minimization of this waste could result in reduction of greenhouse gas emissions of 800e1400 kg/tonne of avoidable food waste. Thus, a minimization strategy would result in increased avoidance of GWP compared to the current situation. The study clearly shows that although modern alternatives for food waste treatment can result in avoidance of GWP through nutrient and energy recovery, food waste prevention yields far greater benefits for GWP compared to both incineration and anaerobic digestion. © 2014 Elsevier Ltd. All rights reserved.

Keywords: Waste minimization Waste reduction Household waste Food waste Life-cycle assessment Carbon footprint

1. Introduction According to the FAO (2011), the amount of food waste generated in the EU equals 280 kg per year for each EU citizen. Of this, 66% is generated in the production to retail chain and 34% by households. Thus, food waste accounts for a large part of the municipal solid waste generated by households. Previous studies have shown that the fraction of food waste in solid household waste equals 38% in Sweden, (IVL, 2002) 50e70% in Brazil (Mahler € et al., 2002), 43% in Turkey (Banar and Ozkan, 2008) and 41% in Denmark (Riber and Christensen, 2006). The European Union Waste Framework Directive (WFD) encourages separate collection and recycling of bio-waste and schemes for source-separation of this fraction have been introduced in several European countries. Due to the energy and nutrient content of this waste and the potential for its recovery in the treatment process, previous studies have suggested that treatment of food waste can result in net environmental benefits using anaerobic digestion or composting alternatives (Møller et al., 2009; Boldrin et al., 2009; Smith et al., 2001; Hirai et al., 2000). The WFD also encourages member states to use life-cycle assessment (LCA) to determine the most

* Corresponding author. E-mail address: [email protected] (A. Bernstad Saraiva Schott). http://dx.doi.org/10.1016/j.jenvman.2014.07.048 0301-4797/© 2014 Elsevier Ltd. All rights reserved.

environmentally beneficial treatment alternative for food waste and other types of bio-waste in the specific local context. The use of LCA as a decision support tool in solid waste management policymaking, as previously proposed by Kirkeby et al. (2006), is therefore likely to increase in the coming years. The current levels of food waste generation in Europe to a large extent derive from mismanagement of edible food (WRAP, 2008; Salhofer et al., 2008). According to the EU waste hierarchy (European Parliament, 2008), prevention should be the main strategy to decrease the environmental burdens from solid waste in member states. However, the focus on LCA of solid waste management systems is commonly related to comparisons of different treatment alternatives for a specific amount of generated solid waste, while potential environmental benefits from waste minimization commonly not are addressed. 1.1. Definitions The EU WFD definition of bio-waste, use the term “food and kitchen waste from households, restaurants, caterers and retail premises, and comparable waste from food processing plants” (European Parliament, 2008). However, in the present study, as well as in many other academic works in this area, the focus is limited to food waste exclusively. Parfitt et al. (2010) makes a distinction between food losses and food waste, where the former

220

A. Bernstad Saraiva Schott, T. Andersson / Journal of Environmental Management 147 (2015) 219e226

is related to losses taking place in production, post-harvest and processing stages of the food supply chain and the latter occurs in the retail and final consumption parts of the chain. However, in the present paper, food waste is discussed only in relation to the very last step of the chain e generated by end-consumers. When discussing food waste prevention, it is important to distinguish between different types of wastes. First, a distinction can be drawn between avoidable and unavoidable food waste. The need to differentiate avoidable and unavoidable food waste has previously been highlighted (WRAP, 2008; Salhofer et al., 2008). Unavoidable food waste can be defined as waste that occurs in the preparation of food: peels, bones, shells etc., which commonly are not are regarded as edible. Avoidable food waste can be defined as products which could have been eaten and consists of prepared but uneaten food (e.g., cooked pasta), food which was left to go bad (e.g., dry bread or rotten fruits and vegetables) and other food products that were disposed of in edible condition. In some cases, a third category e possibly avoidable food waste e has been defined as food waste which in some gastronomic cultures is seen as avoidable, but as unavoidable in others (WRAP, 2009). Some examples are bread crusts and potato peels. In the present study, only the two categories avoidable and unavoidable food waste are used. Based on the definitions above, it can be argued that unavoidable food waste is a result of the very nature of the food we consume at home. If this waste had not occurred in the home as a part of the food preparation process, it would have emerged earlier in the food production chain. Sale of unpeeled and peeled carrots can serve as an example of this. In the first case, a household will produce a larger amount of food waste through the peeling of carrots before consumption. In the latter case, the peeling takes place in industrial facilities and increases the production of food waste from such facilities. The elimination of peels prior to retail sale could also increase the need for packaging and thus result in increased resource utilization and environmental impacts. Such impacts are not considered in the present paper. However, this example clearly demonstrates that in order to address the environmental benefits related to food waste prevention, one must focus on minimizing the avoidable food waste fraction.

“unopened packaging” and “opened packaging” in cases where food was disposed of in its original packaging. Thus, packaging was not separated from the content and assumptions were made in relation to the ratio of packaging to food waste. The sub-categories used for avoidable food waste can be used to describe the waste both in different types of food as well as to give information of the life stage of the food product when discarded. The groups for different food types used were: Meat, Bread, Prepared food, Dairy products, Fruits and vegetables and Other. The lifestage categories used were: Unopened packaging, Opened packaging, Half-eaten food (unprepared left-overs, for example half-eaten apples), Prepared food (food which had been cooked/fried etc. before being discarded, for example cooked pasta or fried meat), Non packaged whole vegetables/fruits (for example whole, uneaten apples), Other meat (unprepared) and Other avoidable food (mostly candy, potato chips and popcorn).

1.2. Aim and scope

2.2. Environmental impact assessment

The present paper reports the potentials for household food waste prevention based on a case study in southern Sweden. An assessment was also made of environmental impacts related to two different treatment alternatives for food waste, both unavoidable and avoidable, by modeling of direct and upstream and downstream impacts related to treatment of the functional unit through anaerobic digestion on the one hand and incineration on the other.

LCA methodology, as described by Finnveden et al. (2009), was used, using system expansion and based on a consequential approach. The avoidable food waste fraction was classified as 100% preventable while the unavoidable food waste fraction was seen as unpreventable. Waste prevention was evaluated through modeling upstream and direct emissions associated with production of avoided food and packaging material. Alternative treatment of this waste was modeled as direct as well as upstream and downstream impacts related to treatment of the functional unit through anaerobic digestion, composting and incineration. The assessment was limited to emissions of greenhouse gases.

2. Methodology 2.1. Waste composition analysis method Three waste composition analyses were performed in a multi€ , southern Sweden. In this area, family residential area in Malmo household food waste has been collected separately in paper bags since 2008. All separately collected food waste and 50% of the bins for disposal of residual waste (randomly selected) were analyzed. n and Lagerkvist This approach is described in detail by Dahle (2008). Waste from a total of 486 households was investigated. The main categories used in the analyses were avoidable and unavoidable food waste. These fractions were divided into a total of eleven sub-fractions, which in some cases were divided even further. Thus a total of 19 fractions were used in the analyses (Table 1). The weight of packaging was included in the categories

Table 1 Sub-fractions used in the detailed assessment of avoidable and unavoidable food waste. Avoidable

Unavoidable

Unopened packaging Meat Other unopened food Opened packaging Meat Bread Dairy products Vegetables and fruits Other opened food Half eaten food Vegetables and fruit Dairy products Prepared food With meat Without meat Non packaged whole vegetables/fruits Non packaged whole bread Other meat Other avoidable food

Tea and coffee grind Peels, shells, cores and trimmings Bones, skin, fat Other unavoidable

2.3. Function unit and system boundaries The functional unit was defined as the service of managing one tonne (metric ton) of food waste from Swedish households. However, waste prevention inherently changes the functional unit (Ekvall et al., 2007). Cleary (2010) uses the terms primary and secondary functional units to ensure both a fixed amount of MSW managed in scenario comparisons including waste prevention, as well as identical reference flows of functionally equivalent product services. However, the same author also states that a secondary functional unit is not required to ensure the functional equivalence of product services if addressing services that are deemed

A. Bernstad Saraiva Schott, T. Andersson / Journal of Environmental Management 147 (2015) 219e226

unwanted by certain segments of the population, such as unsolicited advertising material. Analogously to this, avoidable food waste is in the present study regarded as an unwanted product and secondary functional units were not defined. Similarly to Gentil et al. (2011) it was considered that the quantity of prevented waste is a virtual waste flow. Thus, the consequences of reducing a waste fraction in different waste management systems can be assessed simultaneously as the avoided production impacts from the quantity of prevented food waste measured, without affecting the functional unit (Fig. 1). 2.3.1. System boundaries in waste management systems Direct emissions from transport, pretreatment, treatment and final disposal of secondary waste or use of produced bio-fertilizers were included in the study. Also, the impacts related to production of collection material were addressed. In all cases it was assumed that bio-fertilizers could be used to replace chemical fertilizers. System expansion was used to address energy and nutrient recovery. Marginal data was used for use and substitution of power (0.887 kg CO2-eq/kWh based on Fruergaard et al., 2009) and heat (0.11 kg CO2-eq/MJ based on Gode et al., 2011). Emissions from use of bio-fertilizers on farmland were addressed, while ash treatment not was considered. The collection of input data was restricted to the information that could affect the GWP from compared scenarios. 2.3.2. System boundaries in avoided production systems GWP from prevention of avoidable food waste was assessed through modeling of production, transport (from producer to

1 ton food waste (avoidable+ unavoidable)

Collection and transportation

1 ton food waste (avoidable+ unavoidable)

Collection and transportation

0.65 ton unavoidable food waste

Non-production of 0.35 ton avoidable food waste

3. Life cycle inventory The composition of household food waste was based on a subdivision of avoidable food waste into six categories: Meat, Bread, Prepared food, Dairy products, Fruits and vegetables and Other. Thus, assumptions were made regarding further distribution of these categories in order to collect needed input data in the LCA (Table 2). The evaluation of the avoided production impacts was based on previously performed life cycle assessment of food production (Table 1, SI). Environmental impacts related to food production can vary greatly depending on how (i.e., organic/conventional production) and where the production occurred. Such information commonly cannot be gained through waste composition analyses so assumptions had to be made regarding the origin of the food waste assessed in the study. In order to address these uncertainties throughout the study, two datasets were created: a high-impact (HI) and a low-impact (LI) scenario. The same was done for the consumer transport and preparation processes (Table 2, SI). The prepared food fraction consisted mainly of cooked pasta, potatoes and rice, according to observations during the waste composition

Incineration

Ash treatment

pretreatment

AD

Incineration

Ash treatment

Substitution of chemical fertilizers

Digestate

Substitution of electricity and heat

Collection and transportation

Incineration

Ash treatment

Non-production of 0.35 ton avoidable food waste 0.65 ton unavoidable food waste

retail), final transport (from retail to household) and food preparation, through a production life cycle inventory (LCI). Also, the production of packaging found together with avoidable food waste was included. Emissions related to storage (i.e. refrigeration or freezing of food) was not included in the study, as results from waste composition analyses give very little information relation to such services.

Substitution of electricity and heat

Substitution of car fuel

Substitution of electricity and heat Collection and transportation Substitution of car fuel Substitution of chemical fertilizers

221

pretreatment

AD

Digestate

Incineration

Ash treatment

Substitution of electricity and heat

Fig. 1. Graphical representation of investigated systems.

222

A. Bernstad Saraiva Schott, T. Andersson / Journal of Environmental Management 147 (2015) 219e226

Table 2 Composition of food waste (unavoidable and avoidable) (%) as average based on three waste composition analyses. Type of waste Meat Pig Cow Chicken Bread Dairy Cheese Yoghurt Cream Vegetables and fruit Carrot Leek Tomato Cucumber Lettuce Broccoli (frozen) Apple Orange Melon Prepared food Pasta Rice Potatoes Coffee/Tee Peels etc.a Bones etc.b Other Total a b c d

Avoidable (%)

Unavoidable (%)

10.0 5.0 2.5 2.5 15.0 3.0 2.4 0.3 0.3 37.0 4.1 4.1 4.1 4.1 4.1 4.1 4.1 4.1 4.1 27.0 9.0 9.0 9.0

8.0c 100

12%

8%

5%

12%

Meat Dairy Vegetables Bread Other

63% Fig. 2. Avoidable food waste divided in five different food groups.

4. Results 4.1. Results of waste composition analyses

19.0 66.4 9.5 5.2d 100

Including non-edible parts of vegetables/fruits and egg shells. Including non-edible parts of animals. Including mainly crisps, candy and marmalade. Mostly flowers and tissues.

The results of the detailed analyses in the case study area show average percent of avoidable food waste equaling 35% (SD ¼ 10.5%). The results were divided into life stages and food groups (Figs. 2 and 3). The fractions of dairy products and meat were higher in avoidable food waste found in residual waste. Also, while opened packaging accounted for 42.5% (SD ¼ 10.5%) of the avoidable food waste in residual waste, this fraction accounted for only 24.5% (SD ¼ 9.0%) in of avoidable, separately collected food waste. Unopened packaging was not found among separately collected food waste, while it accounted for 6.6% of the avoidable food waste found in residual waste. 4.2. GWP from food waste management

analyses. An equal division between these food types was assumed. Packaging was included only when food waste was found in its original packaging in the waste composition analysis. According to the analyses, 35% of the avoidable food waste was found in original packaging. However, the analytical method did not take the type of packaging into consideration. Instead, assumptions were made on the fraction of plastic, paper, metal and glass packaging disposed together with avoidable food waste, based on the average division of these materials in household waste from the same area, as previously reported by Bernstad et al. (2012a). The assumptions made in relation to the ratio of packaging in relation to the total amount of packed food disposed of were based on averages from Wallman and Nilsson (2011). It was assumed that packaging would have been incinerated with energy recovery under the same conditions as the food waste incineration process. Data on lower heating value (LHV) and dry matter were obtained from Riber and Christensen (2006) (Table 3, SI). Evaluation of the environmental impacts related to treatment food waste was based on literature data and previously performed life-cycle assessments of food waste management (Table 4, SI). The energy content (as lower heating value, LHV) in food waste has in previous LCA studies has been assumed to range from 1748 to €rjesson and Berglund, 6300 MJ/tonne wet waste (Lee et al., 2007; Bo 2007). 4820 MJ/tonne wet waste was assumed in the present study. Potentials for nutrient recovery, biogas production and energy recovery from total food waste were based on literature values, while the figures for unavoidable food waste were based on primary data (Table 5, SI). GWP impacts related to an avoided production and use of energy and mineral fertilizers recovered through food waste treatment processes were estimated using literature values (Table 6, SI).

The GWP from the two alternative treatments for generated food waste (anaerobic digestion and incineration) is presented in Table 3 divided into different processes in the treatment chain. 4.3. GWP from production, transport and preparation of avoidable food waste The GWP from production, transport and preparation of avoidable food waste is presented in Fig. 4 (high and low assumptions). The results are based on the composition of avoidable food waste and GWP from the production/consumption chain presented above. The results from the combination of prevention of avoidable food waste and management of unavoidable food waste as well as those for treatment of both avoidable and unavoidable food waste are presented in Fig. 5.

6% 32%

14%

Unopened packaging Whole food Opened packaing Halfeaten food

Prepared food

12% 36%

Fig. 3. Avoidable food waste divided on five different life-stages.

A. Bernstad Saraiva Schott, T. Andersson / Journal of Environmental Management 147 (2015) 219e226 Table 3 GWP from food waste management as total food waste (avoidable þ unavoidable) and unavoidable food waste (kg CO2-eq/ton currently generated food waste) divided on processes in the waste management chain. Waste management process

Anaerobic digestion Total food waste

Collection material Collection/transportation Pretreatment Treatment (energy use) Treatment emissions Upgrading (energy use) Methane emissions upgrading Transportation of secondary waste Spreading of biofertilizers Farmland emissions Energy recovery from combustion Avoided fuel production Avoided fertilizer production Total

Incineration

Unavoidable food waste

1.2 7.2 26.6 23.9 7.0 9.1 36.8 6.1

0.7 4.3 16.0 18.2 2.7 6.0 14.7 3.7

0.5 80.6 37.6

0.3 45.6 22.6

288.7 66.8

112.2 16.0

194.1

38.6

Total food waste

Unavoidable food waste

6.8 7.2

4.1 4.3

71 15.5

48 9.3

4.8

188

82.8

3.84

113

43.3

223

4.4. Sensitivity analyses Variations in literature data used for the assessment of GWP from production, transportation and preparation of avoidable food waste were addressed here through the presentation of a high and low impact scenario. Therefore, sensitivity analyses were primarily carried out in relation the waste management alternatives. Environmental impacts from different types of management and treatment alternatives of municipal solid waste can, according to previous studies, vary greatly (Gentil et al., 2010; Morris, 2011). Differences are often related to methodological variations (such as the view on biogenic emissions of CO2 in relation to global warming), differences in system boundary setting (e.g., included/ €rklund, 2002). excluded processes) or variations in input data (Bjo The following factors were chosen for performance of sensitivity analyses, as they have previously been shown to have a large impact on the results in studies of similar systems (Lantz et al., €rjesson and Berglund, 2007; Sonesson 2009; Smith et al., 2001; Bo et al., 2000) (Table 4). In Fig. 6 presents best- and worst-case scenarios for increased GHG emissions from treatment, combined with decreased avoidance through substitution of energy and fertilizer production. The results of the sensitivity analyses show that these changes cause large alterations, primarily in relation to the net GWP from anaerobic digestion of both avoidable and unavoidable food waste, while differences between the base case and the high/low scenarios are

Meat Bread

Low

Dairy Vegetables and fruit Pasta/rice/potatoes Preparing of food

High

Other Packaging 0

200

400

600 800 1000 1200 1400 kg CO2/ton generated food waste

1600

Consumer transports

1800

Fig. 4. GWP from production, transports and preparation of avoidable food waste (kg CO2-eq/ton avoidable food waste) as high and low assumptions.

Incineration Avoided production AD

Waste management

Low avoided + Incineration Low avoided + AD High avoided + Incineration High avoided + AD

-1800

-1600

-1400

-1200 -1000 -800 -600 kg CO2-eq per ton food waste

-400

-200

0

Fig. 5. GWP from non-production of avoidable food waste (high and low assumptions) combined with either anaerobic digestion (AD) or incineration of unavoidable food waste, as well as either anaerobic digestion or incineration of total food waste (avoidable þ unavoidable) as kg CO2-eq/ton currently generated food waste.

224

A. Bernstad Saraiva Schott, T. Andersson / Journal of Environmental Management 147 (2015) 219e226

Table 4 Sensitivity analyses performed in relation to the waste management system. Process

Change

Pretreatment (material losses) Treatment (energy recovery) Treatment (energy use) Methane emissions treatment Upgrading (energy use) Methane emissions upgrading Direct emissions from combustion Energy recovery from combustion Avoided fuel production Avoided fertilizer production

High/low based on Bernstad et al. (2012b) (±) 20% (±) 50% electric and thermal energy (±) 50% methane emission (±) 50% energy use (±) 50% methane emission Excluding emissions of N2O (±) 20% (±) 20% (±) 20%

smaller in relation to incineration alternatives. Also, changes in input data used in assessment of GWP from the waste management system can change the hierarchy between the two compared treatment alternatives. In the case of anaerobic digestion of unavoidable food waste, the results change from net avoidance to net contribution to GWP (Fig. 6). The most influential processes were related to energy input in pretreatment, substitution of fuel, emissions from farmland and fugitive emissions from anaerobic digestion in relation to the biological treatment alternative, as well as assumptions regarding energy recovery and emissions of N2O from incineration. 5. Discussion 5.1. Uncertainties related to LCI data and waste composition analyses As seen in Fig. 5, the net-benefits from food waste minimization vary greatly depending on assumptions related to the GWP from the production of avoidable food waste. In previous studies, Gentil et al. (2011) assumed a GWP of 5.3 kg CO2-eq/kg of meat waste and 1.0 kg CO2-eq/kg of vegetable waste, while Mogensen et al. (2011) used values for avoidable meat production ranging from 5.6 and 5.7 kg CO2-eq/kg of pork and chicken to 32.7 kg CO2-eq/kg of beef, with an average (assuming equal parts of these types of meat) of 14.7 kg CO2-eq/kg. The same author assumes emissions equal to 1.0 CO2-eq/kg for avoidable vegetable waste and 12.2 kg CO2-eq/kg avoidable cheeses. This can be compared to the results from the present study: 8.3e11.8.3 kg CO2-eq/kg of avoidable meat waste, 0.7e1.0 kg CO2-eq/kg of avoidable vegetable waste and 100

All food waste

10.8e13.3 kg CO2-eq/kg of avoidable cheese waste. Thus, the findings of the present study are in agreement with data observed in previous studies. Reported benefits from anaerobic digestion and incineration of food waste are also similar to earlier results (Møller et al., 2009; Smith et al., 2001). Based on assumptions made on the influence of original packaging, food packaging amounted to 4.3% by mass of avoidable food waste, i.e. 50% of the results presented by Lebersorger and Schneider (2011). Sensitivity analyses show that even a doubling of this not would influence the results to any large extent. Thus, according to results from this study, food waste packaging does not have a large influence on the overall GWP related to food waste minimization. 5.2. LCA hot spot identification Identification of hot spots, i.e., factors responsible for a large contribution to the overall GWP from compared scenarios, was done both in relation to the generated avoidable food waste and the compared waste management alternatives. The most critical factors related to the anaerobic digestion waste management alternative are; amount and environmental impacts related to fuels avoided by produced biogas, amount and environmental impacts related to chemical fertilizers avoided by recovered nutrients and emissions from farmland during spreading of digestate. The most critical factors related to the incineration waste management alternative are; amount and environmental impacts related to energy use avoided by incineration, amount and environmental impacts related to energy use in the incineration process and treatment emissions, specifically N2O. Food preparation in the individual household can have a high impact on GWP related to generation of avoidable food waste. In this example, where the fraction of prepared and wasted food represents 27% of the total amount of avoidable food waste, food preparation contributes to almost 20% of the total GWP from the generated avoidable food waste. However, this result is strongly connected to the environmental profile of energy used in the preparation. Thus, assuming a more fossil lean electricity mix (Swedish average, Uppenberg et al., 2001), food preparation would represent around 1% of the total GWP. At the same time, the waste composition analyses on which the data used in the present study were based did not take into consideration that parts of the avoidable food waste in some cases had been prepared industrially before discard. A previous comparison of GHG-emissions from home-

Unavoidable food waste

Base case

0 CO2-eq/ton food waste

Low

High -100

-200

-300

-400

-500

Anaerobic digestion

Incineration

Anaerobic digestion

Incineration

Fig. 6. Results from sensitivity analyses in relation to the waste management alternatives anaerobic digestion and incineration, presented as base case (with the assumptions presented in Table 3) as well as high and low scenarios, according to changes presented in Fig. 4.

A. Bernstad Saraiva Schott, T. Andersson / Journal of Environmental Management 147 (2015) 219e226

cooked, semi-prepared and ready-to-eat food, concluding that semi-produced and ready-to-eat food commonly generate higher emissions compared to home cooking, mainly due to increased need for packaging, storage and inefficient transports (Sonesson, 2005). Thus, it is likely that the GWP related to preparing food that goes to waste is underestimated in the present study. Consumer transports contribute to 2.5% of total GWP from avoidable food waste if assuming that transportations are made in private cars, using diesel. Thus, consumer transport was not of key importance to overall GWP. The results from the study are strongly associated with the type of food wasted by households, as GWP from production of different types of food can vary by several orders of magnitude. Beef, cheese and rice result in especially high emissions of CO2-equivalents due to large emissions of methane in the production chain. Production of 1 kg of beef results in emissions 120 times larger than production of 1 kg of carrots (SIK, 2009a,b). In fact, the results of this study suggest that as much as 22e24% and 7e9% of the GHG-emissions from wasted food are related to the meat and dairy products fractions, while vegetables and fruits correspond to 6e10%. According to Kramer et al. (1999), 28% of Dutch household food consumption is related to meat and 23% to dairy products, while vegetables, fruits and potatoes correspond to 15%. In a similar study of UK households, 30% of total GHG emissions from food consumption were related to meat, 20% to dairy products and 11% to fruits and vegetables (including exotic fruits) (Audsley et al., 2009). Thus, there seems to be similarities in the patterns seen in the GHG emissions from wasted food, where vegetables and fruits correspond to a relatively small part of the total GWP, although the fraction makes up almost 40% of the total mass of generated food waste. Decreasing the amount of specific categories of avoidable waste could therefore have a large impact on the overall savings from reduction of food waste generation. 5.3. Rebound effects related to waste minimization According to Gentil et al. (2011), needs satisfied by products not produced e and thus not ending up as waste e will have to be served by other means. As an example, reduced production of unsolicited paper advertisements will reduce the production of paper waste. However, it could be argued that reduced use of this type of advertising will be compensated by increased use of web-based advertising and one would have to include any environmental impacts connected to this in order to maintain a just comparison between the alternatives. However, in the case of food waste, we do not believe such considerations are justified since it cannot be assumed that a decreased amount of avoidable food waste would cause any increase in environmental burdens in other parts of the system. The only potential rebound effect that is relevant in relation to food waste minimization is related to a general rebound effect when consumers change their behavior, resulting in both environmental and monetary savings. If monetary savings are used to consume more in other areas, environmental savings gained in one € lster, 2008). area could be offset by increased impacts in others (Fo 5.4. Impacts on the waste management chain Waste prevention measures have potential not only to reduce the total amount of food waste, but also can influence the characteristics of remaining food waste. This can have both negative and positive impacts on GWP. A lower content of nutrients in food waste means less potential for substitution of chemical fertilizers in a waste management scenario where nutrients are recovered through anaerobic digestion and use of digestate on farmland. This reduces potential GWP avoidance. However, a lower content of

225

nitrogen in digestate can also reduce risks of emissions of nitrous oxides (N2O) from digestate during storage and spreading on farmland (Lantz et al., 2009) as well as formation of NOx and N2O in incineration of food waste (Smith et al., 2001). The present study was performed in a Swedish context. Thus, alternatives for treatment of generated food waste only include technologies where efficient recovery of energy and nutrients are viable and food waste management can thus be connected to environmental benefits. However, in many countries, management of food waste is related to several negative impacts on the environment, such as fugitive methane emissions from landfills. Thus, the net benefits from food waste minimization are likely to be even larger in cases where the general disposal route is restricted to landfills. 5.5. Effects on other environmental impact categories The present study was limited to assessment of GWP and energy consumption related to food waste minimization. Thus, relevant environmental impacts such as eutrophication, acidification and toxicity were not included. According to the Swedish Environmental Protection Agency, the agricultural sector is responsible for around 40% of the human induced emissions of both nitrogen and phosphorus in Sweden (SEPA, 2008). This shows that food production contributes to many other types of negative environmental impacts than GWP and energy use, again indicating the environmental benefits from prevention of avoidable food waste. 6. Conclusions A case study based on three waste composition analyses among multi-family dwellings in southern Sweden shows that on average 35% of the generated household food waste can be classed as avoidable. Through the use of the life-cycle assessment method, the greenhouse gas emissions from an assumed non-generation of this avoidable food waste could be estimated. It was seen that the reduction in greenhouse gas emissions could reach 800e1400 kg CO2-eq/tonne of avoidable food waste. Currently this avoidable waste in Sweden is mostly incinerated or treated through anaerobic digestion, which also can result in net avoidance of greenhouse gas emissions. However, compared to either anaerobic digestion or incineration of both avoidable and unavoidable food waste, it was estimated that the potential reduction of greenhouse gas emissions could increase 6 and 19 times, respectively, if the generation of avoidable food waste was to be reduced to zero and only unavoidable food waste was to be treated through anaerobic degradation or incineration. The results are to large extent dependent on the composition of what can be seen as avoidable food waste and the date chosen to model the production of this waste. However, as a general indication, a large part of the greenhouse gas emissions related to the production of avoidable food waste comes from animal products such as meat and dairy products. Thus, decreasing the amount of specific categories of avoidable waste could have a large impact on the overall greenhouse gas savings related to food waste minimization. The results also suggest that increased focus should be given to food waste minimization rather than just collection and treatment of waste that already has been generated, and provide quantitative estimates of the climate related benefits from such strategies. Appendix A. Supplementary data Supplementary data associated with this article can be found in the online version, at http://dx.doi.org/10.1016/j.jenvman.2014.07. 048.

226

A. Bernstad Saraiva Schott, T. Andersson / Journal of Environmental Management 147 (2015) 219e226

References Audsley, E., Brander, M., Chatterton, J., Murphy-Bukern, D., Webster, C., Williams, A., 2009. How low can we go? An assessment of greenhouse gas emissions from the UK food system end and the scope to reduce them by 2050. http://assets. wwf.org.uk/downloads/how_low_report_1.pdf. € Banar, M., Ozkan, A., 2008. Characterization of the municipal solid waste in Eskisehir City, Turkey. Environ. Eng. Sci. 25 (8), 1213e1220. Bernstad, A., la Cour Jansen, J., Aspegren, H., 2012a. Local strategies for efficient management of solid household waste e the full-scale Augustenborg experiment. Waste Manage. Res. 30 (2), 200e212. Bernstad, A., Malmqvist, L., Truedsson, C., la Cour Jansen, J., 2012b. Need for improvements in physical pretreatment of source-separated household food waste. Waste Manage. 33 (3), 746e764 (accepted for publication). € rklund, A., 2002. Survey of approaches to improve reliability in LCA. Int. J. LCA 7 Bjo (2), 64e72. Boldrin, A., Andersen, J.K., Møller, J., Christensen, T.H., 2009. Composting and compost utilization: accounting of greenhouse gases and global warming contributions. Waste Manage. Res. 27, 800e812. €rjesson, P., Berglund, M., 2007. Environmental systems analysis of biogas systems Bo e Part II: the environmental impact of replacing various reference systems. Biomass Bioenergy 31 (5), 326e344. Cleary, J., 2010. The incorporation of waste prevention activities into life cycle assessments of municipal solid waste management systems: methodological issues. Int. J. LCA 15 (6), 579e589. n, L., Lagerkvist, A., 2008. Methods for household waste composition studies. Dahle Waste Manage. 28 (7), 1100e1112. € rklund, A., Eriksson, O., Finnveden, G., 2007. What life-cycle Ekvall, T., Assefa, G., Bjo assessment does and does not do in assessments of waste management. Waste Manage. 27 (8), 989e996. European Parliament, 2008. Directive 2008/98/EC of the European Parliament and the council of 19 November 2008 on waste and repealing certain Directives. http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri¼CELEX:32008L0098: EN:NOT. FAO, 2011. Global food losses and food waste. Extent Causes Prev. http://www.fao. org/fileadmin/user_upload/ags/publications/GFL_web.pdf. e, J., Heijungs, R., Hellweg, S., Finnveden, G., Hauschild, M.Z., Ekvall, T., Guine Koehler, A., Pennington, D., Suh, S., 2009. Recent developments in Life Cycle Assessment. J. Environ. Manage. 91 (1), 1e21. €lster, S., 2008. Fo €rva €l till va €rldsundergången (Goodbye Armageddon). Albert Fo €rlag, Stockholm, Sweden (in Swedish). Bonniers Fo Fruergaard, T., Astrup, T., Møller, J., Ekvall, T., Christensen, T.H., 2009. Energy use and recovery in waste management and implications for accounting of greenhouse gases and global warming contributions. Waste Manage. Res. 27 (8), 724e739. Gentil, E.C., Damgaard, A., Hauschild, M., Finnveden, G., 2010. Models for waste life cycle assessment: review of technical assumptions. Waste Manage. 30 (12), 2636e2648. Gentil, E.C., Gallo, D., Christensen, T.H., 2011. Environmental evaluation of municipal waste prevention. Waste Manage. 31, 2371e2379. € € glund, J., Palm, D., 2011. Gode, J., Martinsson, F., Hagberg, L., Oman, A., Ho € faktaboken 2011 e Uppskattade emissionsfaktorer fo € r bra €nslen, el, va €rme Miljo och transporter. Varmeforsk. Report No 1183. 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 31steNov 2nd, 2000, Tsukuba, Japan. 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 Manage. Res. 24, 16e26. Kramer, K.J., Moll, H.C., Nonhebel, S., Wilting, H.C., 1999. Greenhouse gas emissions related to Dutch food consumption. Elsevier Energy Policy 27, 203e216.

€rjesson, P., 2009. Systemoptimerad produktion av forLantz, M., Ekman, A., Bo €- och energioptimerad studie av So €deråsens biodonsgas e En miljo €ggnin (Systems Optimized Production of Vehicle Gas e an gasanla €deråsen Biogas Production Environmental and Energy Assessment of the So Plant). Report 69. Environmental and Energy Systems Studies, Lund University, Lund Sweden (in Swedish). Lebersorger, S., Schneider, F., 2011. Discussion on the methodology for determining food waste in household waste composition studies. Waste Manage. 31 (9e10), 924e1933. 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. Sci. Total Environ. 387 (1e3), 42e53. Mahler, C.F., Araujo, F., Paranhos, R., 2002. Solid Waste Management and Composting. Pollution Aquatic Pollution and Solid Waste Acuarius Publication, Editorial Group/Bio-Rio Foundation (in Portuguese). Mogensen, L., Hermansen, J., Knudsen, M.T., 2011. Madspild i fødevareproduktionen e fra primærproduktion til detailled (Food Waste in the Food Production Chain e from Primary Production to Households) (in Danish). Det Jordbrugsvidenskabeligt Fakultet (DJF), Aarhus University. Møller, J., Boldrin, A., Christensen, T.H., 2009. Anaerobic digestion and digestate use: accounting of greenhouse gases and global warming contribution. Waste Manage. Res. 27 (8), 813e824. Morris, J., 2011. Review of LCAs on Organics Management Methods and Development of an Environmental Hierarchy. Alberta Environment, Edmonton AB, US. Parfitt, J., Barthel, M., Macnaughton, S., 2010. Food waste within food supply chains: quantification and potential for change to 2050. Philos. Trans. R. Soc. B 365, 3065e3081. Riber, C., Christensen, T.H., 2006. Method for fractional solid waste sampling and chemical analysis. Int. J. Environ. Anal. Chem. 87 (5), 321e336. Salhofer, S., Obersteiner, G., Schneider, F., Lebersorger, S., 2008. Potentials for the prevention of municipal solid waste. Waste Manage. 28 (2), 245e259. SEPA, 2008. Konsumtionens klimatpåverkan (The Climate Effects from Consumption). Report No 5903. Swedish Environmental Protection Agency, Stockholm, Sweden (in Swedish). € , A., Sonesson, U., Sund, V., Davis, J. Greenhouse Gas SIK, 2009a. Cederberg, C., Flysjo Emissions from Swedish Consumption of Meat, Milk and Eggs 1990 and 2005. SIK Report No 794, Stockholm, Sweden. €d- kommunikationsunderlag (Climate Impacts SIK, 2009b. Klimatpåverkan från bro from Bread). SIK Report No P80427, Stockholm, Sweden (in Swedish). 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. € ma €ssig ja €mfo € relse av olika s€ €lla en måltid Sonesson, U., 2005. Miljo att att framsta €ttbullar respektive kyckling (hemlagat, halvfabrikat och helfabrikat) med ko (Environmental Comparison of Different Ways of Food Preparation). SIK, Stockholm, Sweden (in Swedish). € rklund, A., Carlsson, M., Dalemo, M., 2000. Environmental and Sonesson, U., Bjo economic analysis of management systems for biodegradable waste. Resour. Conserv. Recyc. 28 (1e2), 29e53. Uppenberg, S., Brandel, M., Lindfors, L.G., Marcus, H.O., Wachtmeister, A., Zetterberg, L., 2001. Environmental Facts for Fuels Part 1. Resources Consumption and Emissions throughout the Lifecycle (in Swedish). Swedish Environmental Institute (IVL), Stockholm, Sweden. €ndning från livsWallman, M., Nilsson, K., 2011. Klimatpåverkan och energianva € rpackningar (Climate Impacts and Energy Use from Food Packaging). medelsfo Report No 18-2011. SIK, Stockholm, Sweden (in Swedish). WRAP, 2008. The Food We Waste, ISBN 1-84405-383-0. Banbury, UK. WRAP, 2009. In: Quested, T., Johnson, H. (Eds.), Household Food and Drink Waste in the UK, ISBN 1-84405-430-6. Banbury, UK.