Incorporating Waste Prevention Activities Into Life Cycle Assessments of Residential Solid Waste Management Systems

Incorporating Waste Prevention Activities into Life Cycle Assessments of Residential Solid Waste Management Systems

Julian Cleary

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy, Graduate Department of Geography, University of Toronto

Incorporating waste prevention activities into life cycle assessments of residential solid waste management systems Doctor of Philosophy, 2012 Julian Cleary Graduate Department of Geography University of Toronto

Abstract The four papers of this dissertation explore themes related to waste prevention, the system boundaries, functional units and scale of life cycle assessments (LCAs) of municipal solid waste (MSW) management, as well as the transparency and consistency of the application of LCA methods. The first paper is a comparative analysis of the methodological choices and transparency of 20 LCAs of MSW that were recently published in peer-reviewed journals, and includes a comparison of their midpoint level impact values using statistical indicators. The second paper proposes a conceptual model, designated WasteMAP (Waste Management And Prevention), for evaluating LCAs of MSW which incorporate waste prevention. In WasteMAP, waste prevention through dematerialization is viewed as analogous to waste treatments so long as it does not affect the functional output (product services) of MSW-generating product systems. Papers 3 and 4 comprise the WasteMAP LCA case study. Paper 3 depicts product LCAs of wine and spirit packaging (conventional, lightweight and refillable, each type generating different quantities of waste) at the scale of the individual package and the municipality. At the municipal scale, the LCAs address impacts from the wine and spirit packaging supplied in the City of Toronto, Canada in 2008, and a waste prevention scenario which substitutes lighter weight and reusable containers. The lowest endpoint level impacts out of the five container types studied were associated with refillable containers and aseptic cartons. Paper 4 addresses the Toronto MSW management system and applies the WasteMAP model to allow for the comparison, on a functionally equivalent basis, of the LCA results of a reference scenario, based on 2008 data, with a scenario incorporating six types of waste prevention activities (prevention of unaddressed advertising mail, disposable plastic bags, newspapers, lightweight and refillable wine and spirit packaging, iii

and yard waste). The findings highlight the benefits of waste prevention, and the relative significance of the decision to account for recycled content when modelling waste prevention. The endpoint level impact assessment results using the ReCiPe and Impact 2002+ evaluation methods are in keeping with the assumption in the waste hierarchy that waste prevention has a superior environmental performance.

Acknowledgements

I wish to thank a number of people for helping me with the design and completion of my research. I greatly appreciate the encouragement, insightful suggestions and help of Virginia Maclaren, my supervisor, as well as my committee members: Miriam Diamond, Danny Harvey and Rodney White. I am also grateful for the enthusiasm and interest of my colleagues and friends at the Geography Department in my research. I am grateful to the City of Toronto Solid Waste Management Services (John Baldry, Bonnie Ballam, Anne Wheatley, Irene Zeppieri), the Liquor Control Board of Ontario (Ian Loadman, and special thanks to Tod Stewart), as well as numerous wineries and distilleries and equipment manufacturers, for participating in my study. I am also appreciative of Dr. Sarah Finklestein’s kind offer of the use of her lab for some of my mass measurements. The Recycling Council of Ontario (RCO) provided me with countless opportunities to meet and discuss waste management issues with many people involved in this fascinating field. I wish to thank Usman Valiante of Corporate Policy Group for his valuable suggestions and for introducing me to the RCO. I would like to extend my thanks and appreciation to the Social Science and Humanities (SSHRC) of Canada and for their financial support in the form of a Canada Graduate Scholarship. I would also like to acknowledge the Geography Graduate Expansion Fund for providing funding for my research. Above all, I would like to express my profound gratitude to my parents for their encouragement, support and insightful contributions that were essential for the successful completion of this dissertation, as well as to Nancy Lo for her forbearance and support.

Acronyms

AC Aseptic Carton

APS Alternate Product System

CSU Conventional Single Use

DPFU Downstream Primary Functional Unit

IC&I Institutional, Industrial and Commercial

ISO International Organization for Standardization

LCA Life Cycle Assessment

LCBO Liquor Control Board of Ontario

LSU Lightweight Single Use

MSW Municipal Solid Waste

PET Polyethylene Terephthalate

RFG Refillable Glass

SFU Secondary Functional Unit

TPS Targeted Product System

UPFU Upstream Primary Functional Unit

WasteMAP Waste Management and Prevention

WPA Waste Prevention Activity

vii

TABLE OF CONTENTS

Abstract ii Acknowledgements iv Acronyms v

INTRODUCTION

1 Introduction 2 1.1 Research objectives 4 1.2 Paper descriptions 5 1.2.1 Life cycle assessments of municipal solid waste management systems: A comparative analysis of selected peer-reviewed literature (Paper 1) 5 1.2.2 The incorporation of waste prevention activities into life cycle assessments of municipal solid waste management systems: Methodological issues (Paper 2) 6 1.2.3 Life cycle assessments of wine and spirit packaging at the product and the municipal scale: A Toronto, Canada case study (Paper 3) 7 1.2.4 Waste prevention and life cycle assessment of residential waste management in Toronto, Canada (Paper 4) 7 1.3 References 8

PAPER 1 Life cycle assessments of municipal solid waste management systems: A comparative analysis of selected peer-reviewed literature

2.1 Introduction 11 2.2 Research objectives 12 2.3 Methodology 13 2.3.1 Selection criteria for LCAs of MSW 13 2.3.2 A common basis for the analysis 14 2.4 Results 14 2.4.1 Study area and scale 14 2.4.2 Goals of the reviewed LCAs 16 2.4.2.1 Comparisons of MSW management systems 16 2.4.3 Functional units 17 2.4.4 System boundaries 18 2.4.4.1 The ‘cradle’ and ‘grave’ of waste 19 2.4.4.2 Life cycle environmental emissions from the production of 20 capital / infrastructure 2.4.4.3 Environmental emissions from MSW transportation 20 2.4.4.4 Selection of energy sources 21 2.4.5 Environmental impacts and LCIA 22 2.4.5.1 Impact categories 23 2.4.5.2 Characterization of impacts 23 2.4.5.3 Single score weighted valuations of impacts 24 viii

2.4.6 Types and sources of data 26 2.4.7 Sensitivity analysis 27 2.4.8 Economic costs of MSW treatment 28 2.5 Comparison of results 29 2.5.1 Acidification potential 32 2.5.2 Global warming potential 33 2.5.3 Net energy use 34 2.6 Discussion 36 2.7 Conclusion 37 2.8 References 38

PAPER 2 The incorporation of waste prevention activities into life cycle assessments of municipal solid waste management systems: Methodological issues

3.1 Introduction 43 3.2 Research objectives 46 3.3 Types of waste prevention activities 46 3.4 Attributional and consequential approaches to waste prevention and LCA 48 3.4.1 The attributional approach 49 3.4.1.1 The traditional product LCA 49 3.4.1.2 The traditional LCA of MSW 50 3.4.2 The consequential LCA 51 3.5 The WasteMAP life cycle assessment 51 3.5.1 System boundary 52 3.5.2 Functional units 54 3.5.2.1 Primary functional unit 55 3.5.2.2 Secondary functional unit 56 3.5.3 Waste flows 57 3.5.4 Environmental emissions 59 3.6 Discussion 61 3.7 Conclusion 63 3.8 References 63

PAPER 3 Life cycle assessments of wine and spirit packaging at the product and the municipal scale: A Toronto, Canada case study

4.1 Introduction 68 4.2 Research objectives and methodology 70 4.2.1 Functional units 71 4.2.2 Packaging scenarios 72 4.2.2.1 LCA scenarios for wine and spirit packaging at the scale of 72 an individual package 4.2.2.2 LCA scenarios for wine/spirit packaging consumption at 72 the municipal scale ix

4.2.3 System boundary 76 4.2.4 Data sources 78 4.2.4.1 Life cycle assessment software, databases and impact 78 assessment tools 4.2.4.2 Questionnaire to wine and spirit producers 79 4.2.4.3 Field and laboratory research 80 4.3 LCA input profile 81 4.4 LCA unit processes 84 4.4.1 Aseptic carton production 84 4.4.2 Glass bottle production 85 4.4.3 Polyethylene terephthalate bottle production 86 4.4.4 Production of secondary packaging 86 4.4.5 Transportation of materials 86 4.4.3.1 Transportation of containers from manufacturer to packager 87 4.4.3.2 Transportation of containers from packager to Toronto 88 4.4.6 Rinsing and filling of containers 90 4.4.7 Reuse of containers 90 4.4.7.1 Transportation of used bottles between retailer and 91 cleaning/refilling facility 4.4.7.2 Washing of used glass bottles 91 4.4.8 Waste management 91 4.4.8.1 Waste collection, transportation and sorting 92 4.4.8.2 Recycling 93 4.4.8.3 Landfilling 94 4.5 Life cycle impact assessment results 94 4.5.1 Individual package scenarios 95 4.5.2 Municipal scale scenarios 99 4.5.2.1 Climate change 99 4.5.2.2 Endpoint level impacts 100 4.6 Sensitivity analysis 105 4.7 Critical review 107 4.8 Discussion 111 4.9 Conclusion 112 4.10 References 113

PAPER 4 Waste prevention and life cycle assessment of residential waste management in Toronto, Canada

5.1 Introduction 121 5.2 Research objective 122 5.3 Methodology 123 5.3.1 Residential waste management scenarios 123 5.3.2 Functional units 128 5.3.3 System boundaries 129 5.3.4 Data sources 131 x

5.4 LCA input profile: Mass balance 134 5.4.1 2008 reference scenario 134 5.4.2 Waste prevention scenario 136 5.5 LCA input profile: Unit processes 138 5.5.1 Upstream processes 139 5.5.1.1 Reduced generation of unaddressed admail (“junk mail”) 139 5.5.1.2 Reuse of disposable plastic shopping bags 140 5.5.1.3 Substitution of newspaper articles available online for those 140 printed on newsprint 5.5.1.4 Substitution of refillable and lightweight wine and spirit 141 containers for conventional containers 5.5.1.5 Grasscycling 142 5.5.2 Downstream processes 142 5.5.2.1 Residential waste collection and transportation 142 5.5.2.2 Sorting 143 5.5.2.3 Biological treatment 144 5.5.2.3 Recycling 145 5.5.2.4 Landfilling 147 5.5.2.5 Reuse of wine and spirit containers 147 5.6 Life cycle impact assessment (LCIA) results 148 5.6.1 Midpoint level comparisons 150 5.6.2 Endpoint level comparisons 153 5.6.2.1 Damage to ecosystem quality 154 5.6.2.2 Damage to human health 157 5.6.2.3 Depletion of natural resources 160 5.6.2.4 WPA comparison 162 5.6.2.5 Comparisons of absolute endpoint impact values 163 5.7 Sensitivity analysis 167 5.8 Critical review 168 5.9 Discussion 175 5.10 Conclusion 177 5.11 References 178

CONCLUSIONS

6.1 Summary of research findings 185 6.1.1 Life cycle assessments of municipal solid waste management systems: 185 A comparative analysis of selected peer-reviewed literature (Paper 1) 6.1.2 The incorporation of waste prevention activities into life cycle assessments of municipal solid waste management systems: 187 Methodological issues (Paper 2) 6.1.3 Life cycle assessments of wine and spirit packaging at the product 188 and the municipal scale: A Toronto, Canada case study (Paper 3) 6.1.4 Waste prevention and life cycle assessment of residential waste management: Toronto, Canada case study (Paper 4) 189 6.2 Policy implications and future research trajectories 190 xi

6.2.1 Transparency and consistency of published LCAs of MSW 191 6.2.2 Comparison of environmental impacts of residential waste 192 management systems that incorporate waste prevention activities 6.2.3 Effects of waste prevention activities on the management of residual 193 waste in the MSW management system 6.3 Concluding remarks 194 6.4 References 195

APPENDICES

A1 Glossary 197

A2 Paper 1 - Supplemental tables 202

A3 Paper 3 205 A3.1 Wine and sprit container, closure and capsule data, assumptions and 206 calculations A3.1.1 Containers 206 A3.1.1.1 Description of 2008 LCBO wine and spirit 206 volume sales datasets A3.1.1.2 Procedure used to exclude wine and spirit coolers 206 from the LCBO datasets A3.1.1.3 Mass of wine and spirit containers, without 207 closure or label, by size and type of container A3.1.1.4 Assumed mass of each type of lightweight single 210 use glass container A3.1.1.5 Container mass equations 210 A3.1.2 Closures 211 A3.1.2.1 Mass of closures, by type 211 A3.1.2.2 Material composition of closures, by type 212 A3.1.2.3 Adoption levels of closures, by type 212 A3.1.2.4 Closure mass equations 214 A3.1.3 Capsules 215 A3.1.3.1 Mass of capsules, by type 215 A3.1.3.2 Material composition of capsules, by type 215 A3.1.3.3 Adoption levels of capsules, by type 216 A3.1.3.4 Capsule mass equations 216 A3.2 Wine/spirit packaging transportation data, assumptions and 217 calculations A3.2.1 Transportation of wine/spirit packages to filling facility 217 A3.2.2 Transportation of packaged wines/spirits from the filling 218 facility to Toronto, Canada in 2008 A3.2.3 Transportation of residential waste 225 A3.3 Unit process descriptions 226 A3.3.1 Selected unit processes 226 A3.3.2 Author-defined unit processes 227 xii

A3.3.2.1 Aseptic carton manufacture 228 A3.3.2.2 Bottle washing 229 A3.3.2.3 Glass production and recycling 230 A3.3.2.4 PET recycling 230 A3.3.3 Electricity production mixes 231 A3.3.4 Recycling and avoided burdens 234 A3.4 Unit process inputs 235 A3.4.1 Individual package scale 235 A3.4.2 Municipal scale 240 A3.5 Midpoint level impacts 243 A3.5.1 Individual package scale 243 A3.5.2 Municipal scale 244 A3.6 Questionnaires and research consent 245 A3.6.1 Set 1 246 A3.6.1.1 Wine companies 246 A3.6.1.2 Spirit companies 256 A3.6.2 Set 2 266 A3.6.2.1 Wine companies 266 A3.6.2.2 Spirit companies 268 A3.7 References 270

A4 Paper 4 271 A4.1 Field research 272 A4.1.1 Admail WPA 272 A4.1.2 Newspaper WPA 272 A4.2 Residential waste calculations 274 A4.2.1 Composition of residual waste 274 A4.2.2 Composition of recyclable waste 274 A4.3 Transportation data, assumptions and calculations 275 A4.3.1 Transportation of newsprint 276 A4.3.2 Residential waste transport 276 A4.4 Unit processes 278 A4.4.1 Selected unit processes 278 A4.4.2 Author-defined unit processes 280 A4.4.2.1 Anaerobic digestion of source-separated organic 280 waste A4.4.2.2 HDPE recycling 281 A4.4.2.3 Recyclables, sorted at MRF, for further 282 treatment/Ontario 2008 A4.4.3 Electricity production mixes 282 A4.4.4 Recycling, biological treatment, and avoided burdens 283 A4.5 Unit process inputs 285 A4.6 Additional WasteMAP LCA results 288 A4.7 References 291

xiii

LIST OF FIGURES

Figure 1 Primary functional units, the inclusion of life cycle emissions from the 18 production of capital/infrastructure, and the inclusion of emissions from MSW transport in the reviewed LCAs of MSW management systems

Figure 2 Impact categories included in the reviewed LCAs of MSW management 24 systems

Figure 3 The inclusion of single score weighted valuations of impacts and 26 explicit sensitivity analyses in the reviewed LCAs of MSW management systems

Figure 4 The inclusion of the economic costs of MSW management activities and 29 the economic values of the environmental impacts of MSW management in the reviewed LCAs of MSW management systems

Figure 5 The median values, interquartile ranges, whiskers and outliers of the 33 acidification potential values for MSW management systems classified into landfilling, mixed treatment and thermal treatment scenarios

Figure 6 The median values, interquartile ranges, whiskers and outliers of the 34 global warming potential values for MSW management systems classified into landfilling, mixed treatment and thermal treatment scenarios

Figure 7 The median values, interquartile ranges, whiskers and outliers of the net 36 energy use values for MSW management systems classified into landfilling, mixed treatment and thermal treatment scenarios

Figure 8 Basic system boundary of the WasteMAP LCA 54

Figure 9 The one litre PET bottle, the one litre tetra prisma aseptic carton, and the 72 one litre conventional single use glass bottle (left to right)

Figure 10 System boundary of the LCA of wine/spirit packaging 77

Figure 11 Mass of wine and spirit containers by container type 82

Figure 12 Country / US state of origin of imported packaged wines (% of 89 volume imported to Ontario in 2008)

Figure 13 Country / US state of origin of imported packaged spirits (% of 89 volume imported to Ontario in 2008)

Figure 14 Impact values for the damage to ecosystem diversity endpoint level 97 indicator, identifying the contribution of each LCA component for individual 1 litre wine and 750 ml spirit packages xiv

Figure 15 Impact values for the human health endpoint level indicator, identifying 98 the contribution of each LCA component for individual 1 litre wine and 750 ml spirit packages

Figure 16 Impact values for the resource depletion endpoint level indicator, 98 identifying the contribution of each LCA component for individual 1 litre wine and 750 ml spirit packages

Figure 17 Impact values for global warming potential using 20, 100 and 500 year 100 time horizons and identifying the contribution of each LCA component for the municipal scale scenarios (IPCC 2007 calculation method employed)

Figure 18 Impact values for the damage to ecosystem diversity endpoint level 101 indicator, identifying the contribution of each LCA component for the municipal scale scenarios

Figure 19 Impact values for the damage to human health endpoint level indicator, 101 identifying the contribution of each LCA component for the municipal scale scenarios

Figure 20 Impact values for the damage to resource availability endpoint level 102 indicator, identifying the contribution of each LCA component for the municipal scale scenarios

Figure 21 Percentage contributions of midpoint level impacts to the damage to 103 ecosystem diversity endpoint level damage indicator values, for the municipal scale scenarios

Figure 22 Percentage contributions of midpoint level impacts to the damage to 104 human health endpoint level damage indicator values, for the municipal scale scenarios

Figure 23 Percentage contributions of midpoint level impacts to the damage to 105 resource availability endpoint level damage indicator values, for the municipal scale scenarios

Figure 24 Flows of waste under the 2008 reference scenario for the life cycle 135 assessment of the City of Toronto’s residential waste management system

Figure 25 Flows of waste under the waste prevention scenario for the life cycle 137 assessment of the City of Toronto’s residential waste management system

Figure 26 Net percent change in the midpoint level environmental impacts from 151 the 2008 reference scenario under the TRACI 2 LCIA method

Figure 27 Net percent change in the midpoint level environmental impacts from 151 the 2008 reference scenario under the Impact 2002+ LCIA method

Figure 28 Net percent change in the midpoint level environmental impacts from 152 the 2008 reference scenario under the ReCiPe (H) LCIA method

Figure 29 Damage to ecosystem quality endpoint level damage indicator values 155 for the life cycle components of the 2008 reference and waste prevention scenarios

Figure 30 Percentage contributions of midpoint level impacts to the damage to 156 ecosystem quality endpoint level damage indicator values, for the 2008 reference and waste prevention scenarios (Impact 2002+)

Figure 31 Percentage contributions of midpoint level impacts to the damage to 156 ecosystem quality endpoint level damage indicator values, for the 2008 reference and waste prevention scenarios (ReCiPe)

Figure 32 Damage to human health endpoint level damage indicator values for 158 the life cycle components of the 2008 reference and waste prevention scenarios

Figure 33 Percentage contributions of midpoint level impacts to the damage to 159 human health endpoint level damage indicator values, for the 2008 reference and waste prevention scenarios (Impact 2002+)

Figure 34 Percentage contributions of midpoint level impacts to the damage to 160 human health endpoint level damage indicator values, for the 2008 reference and waste prevention scenarios (ReCiPe)

Figure 35 Resource availability endpoint level damage indicator values for the 161 life cycle components of the 2008 reference and waste prevention scenarios

Figure 36 Percentage contributions of midpoint level impacts to the Resource 161 Availability endpoint level damage indicator values, for the 2008 reference and waste prevention scenarios (Impact 2002+)

Figure 37 Percentage contributions of midpoint level impacts to the Resource 162 Availability endpoint level damage indicator values, for the 2008 reference and waste prevention scenarios (ReCiPe)

Figure 38 Life cycle components of the 2008 reference scenario as a percent of 165 the sum of the absolute values for the endpoint level damage indicators

Figure 39 Life cycle components of the waste prevention scenario as a percent of 166 the sum of the absolute values for the endpoint level damage indicators

xvi

LIST OF TABLES

Table 1 Study area and scale of the reviewed life cycle assessments of municipal 15 solid waste management systems

Table 2 Acidification potential (AP), global warming potential (GWP), and net 30 energy use (NEU) values for each MSW treatment scenario

Table 3 The properties of each type of waste prevention activity 48

Table 4 Estimated consumption of wines and spirits in Toronto by size and type 73 of container, 2008

Table 5 Estimated consumption of non-bag-in-box wines and spirits in Toronto 74 by size and type of container in the alternative packaging scenario

Table 6 Mass of wine and spirit packaging material inputs and waste under the 83 2008 reference and alternative packaging scenarios

Table 7 Average distances travelled by wine/spirit containers from the 87 manufacturer to the packager and from the packager to Toronto in 2008

Table 8 Net endpoint level impacts from the life cycles of the alternative 95 packaging systems as percentages of the net impacts of the CSU glass packaging system

Table 9 Values used in the sensitivity analysis to depict the distances travelled 106 by wine/spirit containers from the manufacturer to the packager and from the packager to Toronto in 2008

Table 10 Percent change from reference net endpoint impact values for individual 106 wine containers

Table 11 Percent change from reference net endpoint impact values for 2008 107 reference and alternative packaging scenarios

Table 12 The waste prevention activities addressed in the waste prevention 124 scenario, the properties of each type of WPA and the quantities of waste prevention

Table 13 Residential waste generation in the City of Toronto, by material 136 composition, 2008

Table 14 Waste flows of each waste management scenario 137

xvii

Table 15 Percent diversion of each residential waste stream and reprocessing 146 efficiency at recyclers and biological treatment facilities

Table 16 Net endpoint level impacts from the 2008 reference and waste 154 prevention scenarios

Table 17 Net endpoint level impacts of each WPA within the waste prevention 163 scenario (Impact 2002+)

Table 18 Net endpoint level impacts of each WPA within the waste prevention 163 scenario (ReCiPe (H))

Table 19 Percentage changes in the endpoint level net avoided impacts from the 167 default WPAs to those in an altered scenario in which the materials subject to waste prevention contain no recycled content

Table 20 Percentage changes in the endpoint level waste transportation impacts 168 from the 2008 reference scenario to an altered scenario in which all landfilled waste is sent to the Green Lane landfill in St. Thomas, Ontario

xviii

Appendix Tables

Table A2.1 Primary functional units and system boundaries of reviewed LCAs of 203 MSW

Table A2.2 Impact categories addressed in life cycle impact assessments and 203 weighted valuations of impacts

Table A2.3 Economic impacts, sensitivity analysis, use of LCA computer 204 models, and electricity networks addressed

Table A3.1 Estimated mass of wine and spirit containers, without closure or 207 label,by size and type of container

Table A3.2 Measured densities of popular spirits purchased in Ontario, used as 208 inputs for calculating the average mass of spirit containers

Table A3.3 Spirit density substitutions for brands with unmeasured densities 209

Table A3.4 Assumed mass of lightweight single use glass bottles 210

Table A3.5 Estimated mass of wine and spirit container closures 211

Table A3.6 Estimated material composition of closures 212

Table A3.7 Estimated adoption levels of container closures for wines and 213 spirits sold by the LCBO in Ontario, by size and type of container, 2008

Table A3.8 Estimated mass of wine and spirit container capsules 215

Table A3.9 Estimated material composition of capsules 216

Table A3.10 Estimated adoption levels of wine container capsules for wine sold 216 in Ontario, 2008

Table A3.11 Wine/spirit import markets and the percent of wine/spirit imports 218 to Ontario, Canada from each in 2008

Table A3.12 HS Codes selected to provide the quantity of packaged wine 220 imported to Ontario, Canada in 2008

Table A3.13 HS Codes selected to provide the quantity of packaged spirits 221 imported to Ontario, Canada in 2008

xix

Table A3.14 List of overseas wine/spirit import markets, the assumed port of 222 export and port of arrival, the shipping distances between these ports, and the respective volumes of packaged wine and spirits shipped to Toronto in 2008

Table A3.15 List of wine/spirit import markets, the land-based shipping distances 224 to the City of Toronto, and the respective volumes of packaged wines and spirits shipped to the city in 2008

Table A3.16 Residential waste treatment/disposal locations and estimated 225 distances from Toronto, mass shipped and tonne-km calculated for shipments to treatment/disposal locations in 2008

Table A3.17 Unit processes for wine/spirit packaging LCAs 226

Table A3.18 Percentage composition of the electricity production mixes of 232 Ontario and wine and spirit import markets

Table A3.19 US-EI unit processes for electricity production categories, and the 233 average electricity production mix for the upstream component of each LCA scenario

Table A3.20 Unit processes for the recycling component of the wine/spirit 234 packaging LCA

Table A3.21 Upstream unit processes and the quantities required under each one 235 litre wine packaging LCA scenario

Table A3.22 Upstream unit processes and the quantities required under each 237 750 ml spirit packaging LCA scenario

Table A3.23 Downstream unit processes and the quantities required under each 238 one litre wine packaging LCA scenario

Table A3.24 Downstream unit processes and the quantities required under each 239 750 ml spirit packaging LCA scenario

Table A3.25 Upstream unit processes and the quantities required under each 240 municipal scale wine/spirit packaging LCA scenario

Table A3.26 Downstream unit processes and the quantities required under each 242 municipal scale wine/spirit packaging LCA scenario

Table A3.27 Midpoint level environmental impacts from the individual wine 243 packaging scenarios under the ReCiPe (H) LCIA method

Table A3.28 Midpoint level environmental impacts from the individual spirit 244 packaging scenarios under the ReCiPe (H) LCIA method

Table A3.29 Endpoint level environmental impacts from the 2008 reference and 244 alternative packaging scenarios under the ReCiPe (H) LCIA method

Table A4.1 Mass of unaddressed admail received at a single detached Toronto 272 household during the week of 7-13 February 2011

Table A4.2 Measured mass of daily subscription newspapers in Toronto during 273 the week of 17-23 January 2011

Table A4.3 2008 daily circulation of subscription newspapers in Toronto based 273 on statistics from the Canadian Newspaper Association (2009)

Table A4.4 Residential waste treatment/disposal locations and estimated 276 distances from Toronto, and tonne-km calculated for shipments to treatment/disposal locations in 2008

Table A4.5 Unit processes for upstream component of the WasteMAP LCA 278 scenarios

Table A4.6 Processes for downstream component of the WasteMAP LCA 279 scenarios (excluding recycling)

Table A4.7 Unit processes for the recycling component of the WasteMAP LCA 284 scenarios

Table A4.8 Upstream unit processes and the quantities required under the waste 285 prevention scenario

Table A4.9 Downstream unit processes and the quantities required under each 286 WasteMAP scenario

Table A4.10 Midpoint level environmental impacts from the 2008 reference and 288 waste prevention scenarios under the Impact 2002+ LCIA method

Table A4.11 Midpoint level environmental impacts from the 2008 reference and 288 waste prevention scenarios under the ReCiPe (H) LCIA method

Table A4.12 Midpoint level environmental impacts from the 2008 reference and 288 waste prevention scenarios under the TRACI 2 LCIA method

Table A4.13 Endpoint level environmental impacts from the 2008 reference and 289 waste prevention scenarios under the Impact 2002+ LCIA method

xxi

Table A4.14 Endpoint level environmental impacts from the 2008 reference and 289 waste prevention scenarios under the ReCiPe (H) LCIA method

Table A4.15 Identification of the most important process contributors to impacts 289 and avoided burdens for each midpoint level category

INTRODUCTION

1 Introduction

How can one evaluate the environmental burdens of a waste management system that incorporates waste prevention activities (WPAs)? This is the primary research question of interest in this dissertation. Rapidly increasing per capita waste generation in developed countries (OECD 2007) has resulted in waste management policies emphasizing waste prevention as an important objective (e.g., Commission of the European Communities 2003). However, there is a lack of evidence from life cycle assessments (LCAs) of waste which addresses the environmental benefits of waste prevention activities. The lack of evidence is a consequence of the exclusion of waste prevention impacts from the system boundary of traditional LCAs of waste, which are used to evaluate the environmental impacts of waste management systems. This dissertation responds to the following claim of Ekvall et al. (2007): “LCA models that calculate the environmental burdens per kg or tonne of waste generated allow for environmental comparisons of different options for dealing with this waste, but not for analyses of changes in the quantities of waste generated. They are inadequate for the identification and assessment of waste prevention strategies.” In a 2009 editorial, the International Journal of Life Cycle Assessment specifically identified the incorporation of waste prevention into LCAs of MSW as an area in which additional research and methodological refinement are required (Gheewala 2009). The following four-paper dissertation provides (1) a comparative analysis of the scope and results of published life cycle assessments (LCAs) of waste; (2) a proposed conceptual LCA model, developed by the author, called WasteMAP, which can evaluate the environmental burdens of a waste management system that incorporates WPAs ; (3) a comparison of the results of product LCAs undertaken at the scale of the individual product and at the municipal scale; and (4) a case study using the WasteMAP model with a comparison of functionally equivalent LCAs of waste management systems, one of which incorporates waste prevention activities. LCA is the method of environmental assessment used to identify and quantify the environmental exchanges, and damages caused by those processes occurring within the system boundary of the life cycle of interest. While the initial development of LCA 3

began as far back as 1969 in a study commissioned by the Coca-Cola Company (Hunt and Franklin 1996), interest in this topic within corporate, government, and academic circles became widespread only in the 1990s, with an academic journal specifically devoted to this subject launched in 1996 (International Journal of Life Cycle Assessment). LCA has since become an important means to help ensure that actions taken to reduce environmental impacts at one stage of the life cycle do not result in upstream or downstream consequences which would produce a negative outcome overall. LCAs are commonly applied by policy-makers, corporations, environmental non- governmental organizations, private consultants, and academics. They have been used to inform environmental product declarations (EPDs) (e.g., ISO 2006a), government and corporate policies and independent critical assessments. LCAs can also contribute information to influence product and system designs, including the designs of waste management systems. Although academic LCAs are commonly undertaken by researchers in engineering departments, many attributes of LCA make it a significant tool for geographical analysis. The scope of an LCA is limited by its geographic circumstances, be it relative to the geography of the technologies required for the processes within the life cycle, and/or the context in which the results are to be interpreted. Relative to their defined goals and scope, product LCAs tend to be less directly tied to their geographical context than are LCAs of waste management systems. With product LCAs, functional units tend to depict the amount of product service supplied, often irrespective of geography, whereas functional units of LCAs of waste tend to be defined by the geographic origin of the waste to be managed. The procedure to undertake an attributional LCA has been described in detail by the International Organization for Standardization (ISO 2006b), among other sources. The attributional LCA, considered the standard form of LCA, is used to evaluate the environmental burdens attributed to each unit process within the defined system boundary of the LCA. More recently, consequential LCAs have been proposed to evaluate the life cycle consequences of a change in the production of a good or service, with no pre-set system boundary. The lack of a prescribed boundary can make 4 consequential LCAs more complicated to undertake.. Complete definitions of the attributional and consequential LCA are located in Appendix 1. LCAs have been designed based on economic input-output tables, such as the EIO-LCA model of Carnegie Mellon University (Hendrickson et al. 2006). Input-output LCA models have more comprehensive system boundaries than attributional LCAs, since they account for all the direct, indirect and induced environmental emissions from each economic sector. Negative aspects of an input-output LCA include its low resolution relative to the economic sector under study, and its linear modelling design (Rebitzer et al. 2004). This dissertation focuses on the more commonly applied attributional LCA. Due to the wide latitude available when defining the goal, scope, and data requirements of a life cycle assessment, it can be difficult for reviewers of LCAs to interpret and compare the results. Therefore, the first paper in this dissertation presents a comparative analysis of recently published LCAs of multi-material waste management systems in order to clarify the types of methodological assumptions selected for each LCA, the consistency of the assumptions, and the similarities in the results. The findings of the research in the comparative analysis pertaining to waste prevention and the “waste hierarchy” helped to inspire the subjects explored in the second, third and fourth papers. The second paper examines the methodological issues associated with the incorporation of waste prevention activities into LCAs of waste. It also proposes a conceptual model, in keeping with ISO standards for LCA, to permit functionally equivalent comparisons of the results of waste management scenarios with varying levels of waste prevention. The third paper consists of evaluations of the environmental impacts of wine and spirit packaging using product LCAs applied both at the municipal scale (specific location: City of Toronto, Canada) and at the scale of the individual package. Paper 4 is a case study of the proposed conceptual model from Paper 2, using the residential waste management system of the City of Toronto in tandem with waste prevention activities for numerous waste streams.

1.1 Research objectives

This dissertation has the following objectives: 5

(1) To quantify and compare the current methodological assumptions and results of LCAs of municipal solid waste management systems recently (2002-2008) published in peer- reviewed journals;

(2) To propose a conceptual LCA model for evaluating and comparing, on a functionally equivalent basis, waste management systems that address different quantities of waste;

(3) To produce LCAs of packaging systems (responsible for varying levels of waste generation) at the scale of the individual package and at the municipal scale, and to identify the ways in which the results of these LCAs differ; and

(4) To evaluate and compare the results of a traditional LCA of MSW and an LCA scenario incorporating waste prevention activities, using the conceptual Waste Management And Prevention (WasteMAP) LCA model.

1.2 Paper descriptions

1.2.1 Life cycle assessments of municipal solid waste management systems: A comparative analysis of selected peer-reviewed literature (Paper 1)

While reading the LCA literature on MSW management systems, it became increasingly evident that it would be very difficult to draw conclusions based on the results. This difficulty arose because so many published LCAs do not make clear their system boundaries and assumptions. While there have been a number of studies comparing the results of LCAs of MSW, a comparative analysis of the methodological transparency of published LCAs had not been pursued. I felt it was necessary to undertake such a study in order to highlight the extent of the problem, hopefully leading to LCAs whose results would be easier to compare and contrast. The primary objective of the first paper of the dissertation, published in Environment International in November 2009 (Cleary 2009), was to provide an indication of the frequency of use of particular methodological assumptions made during 6

the goal and scope definition phase of LCA, as well as the relative clarity of LCA practitioners in revealing what they are. Paper 1 addresses the goal, scope and results of 20 LCAs of mixed-material MSW management systems published between 2002 and 2008 in 11 English-language peer-reviewed journals. The findings are used to help identify the areas of both agreement and disagreement among researchers in defining the scope of an LCA for the study of MSW management systems. This comparative analysis quantifies the methodological transparency of the studies and the frequency of use of particular system boundaries, types of data sources, environmental impact categories, impact weightings, economic valuations, and sensitivity analyses. Net energy use (NEU), global warming potential (GWP), and acidification potential (AP) values for various types of MSW management systems are also compared using statistical indicators. This paper addresses the first of the research objectives.

1.2.2 The incorporation of waste prevention activities into life cycle assessments of municipal solid waste management systems: Methodological issues (Paper 2)

In the comparative analysis of the first paper, I found it a common practice for the authors of published LCAs of waste to claim that they are evaluating the validity of the waste hierarchy, yet these LCAs omit the first two components of this hierarchy – waste prevention and product reuse. The second paper of this dissertation, which was published in the International Journal of Life Cycle Assessment in July 2010 (Cleary 2010), proposes a conceptual model to facilitate the use of LCAs of waste to pursue such an evaluation of the waste hierarchy, and incorporate waste prevention activities (WPAs). The second paper introduces my Waste Management and Prevention (WasteMAP) LCA, a conceptual model that applies system expansion to generate a hybrid of the traditional product and waste LCA. WasteMAP, unlike the traditional LCA of MSW, compares functionally equivalent MSW management scenarios incorporating both treatment and prevention. Integral to the WasteMAP proposal is an acceptance of the apparent philosophical leap required to consider waste prevention a form of waste management. This paper addresses the second research objective.

1.2.3 Life cycle assessments of wine and spirit packaging at the product and the municipal scale: A Toronto, Canada case study (Paper 3)

Several types of alternative packaging systems for wines and spirits have been introduced recently in Ontario and promoted as methods to reduce the quantity of residential waste generated, and the associated environmental impacts (LCBO 2006). Since waste prevention in the WasteMAP LCA applies to dematerialization through substitution, wine and spirit packaging alternatives to conventional single use glass bottles are appropriate examples of WPAs for the waste management scenario of the WasteMAP LCA case study in Paper 4. Paper 3 provides results from product LCAs undertaken at the scale of an individual package, as well as at a municipal scale. LCAs at the former scale address the life cycles of one litre wine containers and 750 ml spirit containers. The types of containers evaluated include aseptic cartons (for wine only), PET plastic containers, conventional and lightweight single use glass bottles, and refillable glass bottles. At the municipal scale, the system boundary of the LCA is adapted to encompass the life cycles of all of the glass, plastic and aseptic carton packaging of wines and spirits supplied to residents of the City of Toronto in 2008. The results from this reference scenario are compared to those of a scenario comprising higher levels of alternative packaging (i.e., packaging other than conventional single use glass bottles). The LCAs in Paper 3 do not exemplify the use of the WasteMAP system boundaries and functional units. However, they do illustrate the application of product LCAs at a municipal scale, an intermediate step that is required for the depiction of waste prevention activities within WasteMAP LCA system boundaries. Some of the data inputs for the LCAs at the municipal scale are used for the larger WasteMAP LCA undertaken in Paper 4.

1.2.4 Waste prevention and life cycle assessment of residential waste management in Toronto, Canada (Paper 4)

Paper 4 is a case study of the conceptual WasteMAP LCA model defined in Paper 2, using some of the data obtained for Paper 3. It includes a comparison of the net environmental burdens of the 2008 reference scenario of waste management in Toronto with those of a waste management scenario that incorporates six types of waste prevention activities. The WPAs included within the waste prevention scenario address significant portions of the following waste streams: unaddressed advertising mail, disposable carry out shopping bags, newspapers, wine and spirit packaging, and grass (yard waste). The residential waste management system of the City of Toronto, Canada is selected for this case study due to: (1) availability of detailed data; and (2) the City has a waste diversion target which is deemed to include the effects of some waste prevention activities, such as grasscycling (City of Toronto 2007).

1.3 References

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):579-589.

Cleary J. 2009. Life cycle assessments of municipal solid waste management systems: A comparative analysis of selected peer-reviewed literature. Environ Int 35(8):1256-1266.

Commission of the European Communities. 2003. Towards a thematic strategy on the prevention and recycling of waste. Communication from the Commission. Brussels: Commission of the European Communities. COM(2003) 301 final. Available at: http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=COM:2003:0301:FIN:EN:PDF. Accessed on 2011 05 05.

Cox J, Giorgi S, Sharp V, Strange K, Wilson D, Blakey N. 2010. Household waste prevention – a review of evidence. Waste Manage Res 28:193-219

Ekvall T, Assefa G, Bjorklund A, Eriksson O, Finnveden G. 2007. What life-cycle assessment does and does not do in assessments of waste management. Waste Manage 27(8):989-996

Gheewala S. 2009. LCA of waste management systems—research opportunities. Int J LCA 14:589-590

Heijungs, R; Guinee, J.B. 2007. Allocation and 'what-if' scenarios in life cycle assessment of waste management systems. Waste Manage 27(8): 997-1005.

Hendrickson, C.T, Lave, L.B. and H.S. Matthews. 2006. Environmental Life Cycle Assessment of Goods and Services: An Input-Output Approach. Washington, D.C.: Resources for the Future. 262 pp.

Hunt R, Franklin W. 1996. LCA - How it Came About - Personal Reflections on the Origin and Development LCA in the USA. Int J LCA 1(1):4-7.

ISO. 2006a. Environmental labels and declarations -- Type III environmental declarations -- Principles and procedures. ISO 14025: 2006

ISO. 2006b. Environmental Management – Life Cycle Assessment. Principles and framework. ISO 14040: 2006

Liquor Control Board of Ontario. 2006. LCBO Annual Report 2005-2006. Toronto, Ontario: LCBO Corporate Communications. 46 pp.

Rebitzer G, Ekvall T, Frischknecht R, Hunkeler D, Norris G, Rydberg T, et al. 2004. Life cycle assessment part 1: framework, goal and scope definition, inventory analysis, and applications. Environ Int 30(5):701–720

PAPER 1

Life cycle assessments of municipal solid waste management systems:

A comparative analysis of selected peer-reviewed literature

2.1 Introduction

The popularity of life cycle assessments (LCAs) in analyzing municipal solid waste (MSW) management systems is illustrated by the numerous published studies of the life cycle emissions of these systems, as well as by the substantial number of LCA computer models and databases addressing MSW management. Over the past few decades, many academics, as well as organizations such as the International Organization of Standardization (ISO) and the Society of Environmental Toxicology and Chemistry (SETAC) have contributed to the development of the methodology used to undertake LCAs. Of recent importance is the revision of the requirements and guidelines for LCAs by the ISO in 2006. Comparisons of the results of different LCAs of MSW management systems have been published, with most tending to focus on results from a material-specific basis (Björklund and Finnveden 2005; Finnveden and Ekvall 1998; WRAP 2006). However, comparative studies focused on the methodological transparency of published LCAs of MSW are not undertaken because transparency is perceived as a criterion to filter out studies from review (e.g.,WRAP, 2006; Björklund and Finnveden, 2005), instead of a subject for review. The results from Winkler and Bilitewski (2007) and Winkler (2005) underline the necessity of clearly identifying both the scope and methodological assumptions of LCAs in order to have confidence in the results. In their comparative analyses of six prominent LCA models, both Winkler (2005) and Winkler and Bilitewski (2007) use identical input data for landfilling, incineration and material recovery scenarios for the waste management system of Dresden, Germany. They find very high variations in the predicted emissions, especially for those associated with toxicity, such as heavy metals. The authors suggest that the transparency of the LCA models needs much improvement, although the identification of the specific differences between each model is not part of the research objectives of their important papers. A number of published comparisons of the results of LCAs of MSW management systems have been undertaken from a material-specific perspective. The Waste & Resources Action Programme (WRAP) (2006), in its international review of 55 ‘state-of- the-art’ LCAs designed to evaluate the environmental benefits of recycling, observes a 12

high level of agreement concerning the life cycle environmental emissions from the treatment of recyclable materials. WRAP (2006) indicates that recycling generates greater net environmental benefits than landfilling and thermal treatment in most of the reviewed LCAs, for most of the examined materials. Björklund and Finnveden (2005), Denison (1996), and Finnveden and Ekvall (1998) also review LCA results for MSW management systems and produce results similar to WRAP (2006), while using much smaller sample sizes. Björklund and Finnveden (2005) focus on the material-specific global warming impact and total energy use of MSW management scenarios from 10 LCAs. Denison (1996) examines four North American LCAs of MSW from both a mixed-waste and material-specific perspective. In contrast to the other papers, Finnveden and Ekvall (1998) compare seven sets of LCA results that pertain only to the treatment of waste paper. A comprehensive comparative study of LCAs of MSW that provides the relative frequencies of applying the particular guidelines and requirements associated with the goal and scope of LCA, as detailed in ISO 14044 (2006), has not yet been undertaken, although WRAP (2006) and Björklund and Finnveden (2005) address some system boundary issues. In addition, statistical indicators have not yet been used to compare the results of published LCAs of MSW management systems.

2.2 Research objectives

This study touches upon many of the issues addressed by the aforementioned comparative studies, and is the most comprehensive comparative study of LCAs of mixed-waste MSW management systems yet published in a peer-reviewed journal. It addresses the frequency of use of particular methodological assumptions made during the goal and scope definition phase of LCA, as well as the relative clarity of LCA practitioners in revealing what they are. The findings should help to identify the areas of agreement and disagreement among researchers in defining the scope of an LCA for the study of MSW management systems. As part of this review of LCAs of MSW, the estimates of the net energy use (NEU), acidification potential (AP) and global warming potential (GWP) of each 13

scenario are evaluated using statistical analysis. Other impact categories are excluded from this comparison due to the low number of LCAs that provide the required values (i.e., all of the excluded categories had less than 20 scenarios). A comparison of the relative environmental performances of waste treatment scenario types within studies is also undertaken. Unlike all published result comparisons but Denison (1996), these comparisons investigate the environmental performances of various types of MSW management systems from a mixed-waste perspective. Despite the potential for substantial differences in the methodological assumptions of each of the reviewed LCAs, I believe that this type of comparison can be used to evaluate the preferential order of the waste treatment methods depicted by the waste hierarchy. It can also provide insight into the overall sensitivity of this order to the varying assumptions, system boundaries and system profiles of MSW treatment scenarios, which are classified into thermal, mixed and landfill treatment systems. The waste hierarchy broadly depicts the options to manage waste, by order of preference (Price and Joseph 2000).

2.3 Methodology

2.3.1 Selection criteria for LCAs of MSW

A total of 20 process-based life cycle assessments of MSW management systems published in English between 2002 and 2008 are used in this comparative analysis (Table 1). All of these LCAs pertain to waste management systems that address residential or commercial (non-industrial) solid waste, and include, at a minimum, the plastic, paper, and organic streams. Those LCAs that focus on only one MSW material type, such as organic waste, have been excluded from the review, although one LCA (Arena et al. 2003) addresses the residential waste remaining after households have source-separated much of the recyclable material. Each LCA also addresses a wide array of environmental emissions and/or impacts. Although LCAs are commonly undertaken in the governmental, non-governmental, industrial and consulting sectors, this paper, unlike the detailed WRAP (2006) report, focuses only on academic literature published in peer- reviewed scientific journals. 14

The LCAs selected for my review, three of which are divided into two articles each, comprise 23 articles, and have been published in a total of 11 different journals. Most of the LCAs were selected through a keyword search of “life cycle and waste” in the Web of Science and GEOBASE databases, as well as an online search through Google Scholar. Some were located in the reference sections of published LCAs of MSW. I believe that this sample represents the majority of those LCAs that meet the selection criteria, and therefore, is deemed fairly representative.

2.3.2 A common basis for the analysis

Numerous issues are examined in this comparative analysis, including: (1) study area and scale; (2) goals of the reviewed LCAs; (3) functional units; (4) system boundaries; (5) types of data sources; (6) environmental impacts; (7) sensitivity analysis; (8) economic costs of MSW treatment and (9) the quantitative results for net energy use (NEU), global warming potential (GWP) and acidification potential (AP). Some topics are reviewed relative to their presence or absence in each reviewed LCA (e.g., the presence of an explicit sensitivity analysis). In other cases, the review focuses on how the articles incorporate these subjects into the LCAs. Many of the comparisons provide an indication of the types of results that LCA practitioners feel are important to include in their studies. Most of these topics (1–7) are cited as important in the ISO 14044 Life Cycle Assessment Requirements and Guidelines international standard from 2006. They depict characteristics of LCA design essential to the interpretation of the results. For example, the decision to include the secondary environmental burdens from capital equipment in an LCA would necessarily affect the environmental impacts of all waste management scenarios studied. Some topics, such as the economic costs of MSW treatment, are selected because their significance is debated in peer-reviewed literature addressing LCA.

2.4 Results

2.4.1 Study area and scale 15

The vast majority (17/20) of the reviewed LCAs are undertaken at regional and municipal scales (Table 1). Of the remainder, Finnveden et al. (2005) account for waste treatment in all of Sweden, Chaya and Gheewala (2007) carry out the LCA at the scale of an individual waste treatment facility, whereas Aye and Widjaya (2006) address the waste generated from traditional markets in Jakarta, Indonesia. Some of the LCAs include more than one city. Twelve of the case studies focus on Europe; five are based in Asia, two in North America, and one in South America. Two of the LCAs address hypothetical cities, and use domestic (i.e., Italy and the United States) waste composition and other data for the emission and impact calculations. In these LCAs, the geographic specificity and design of the selected waste management systems tend to be less important than the evaluation of a model, a technical process or a philosophical hypothesis (i.e., the ‘hierarchy of waste’).

Table 1 Study area and scale of the reviewed life cycle assessments of municipal solid waste management systems

Reviewed LCA Study area Scale Arena et al. 2003 Regione Campania, Italy Regional Aye and Widjaya 2006 Jakarta, Indonesia Traditional Markets Beigl and Salhofer 2004 Rural communities in two districts in the Municipal province of Salzburg, Austria Buttol et al. 2007 Bologna District, Italy Regional Carlsson Reich 2005 Uppsala, Stockholm, and Alvdalen, Municipal Sweden Chaya and Gheewala Thailand Waste treatment facility 2007 Consonni et al. 2005 [a, Hypothetical Italian cities of 200 000 and Municipal b] 1.2 million people Di Maria and Fantozzi Umbria region, Italy Regional 2004 Emery et al. 2007 Rondda Cynon Taf County, South Wales, Regional U.K. Eriksson et al. 2005 Uppsala, Stockholm, and Alvdalen, Municipal Sweden Finnveden et al. 2005 / Sweden National Moberg et al. 2005 Hong et al. 2006 Pudong, China Municipal Kirkeby et al. 2006 Aarhus, Denmark Municipal Mendes et al. 2004 Sao Paulo, Brazil Municipal Morris 2005 4 regions in Washington state and San Regional Luis Obispo county in California, U.S. Özeler et al. 2006 Ankara, Turkey Municipal Rodriguez-Iglesias et al. Asturias, Spain Municipal 2003 Shmelev and Powell Gloucestershire, U.K. Regional 2006 Solano et al. 2002 [a, b] Hypothetical U.S. city of 600 000 people Municipal Tan and Khoo 2006 Singapore Municipal

2.4.2 Goals of the reviewed LCAs

The first component of the goal and scope definition phase is the depiction of the goal of the LCA, including whether or not the results will be used for comparative assertions (ISO 2006). The reviewed LCAs commonly evaluate the environmental performance of different MSW management scenarios or provide the quantitative results used to facilitate such an evaluation, and all of them make comparative assertions regarding the scenarios. Their objectives vary from the general (e.g., testing the validity of the ‘hierarchy of waste’), to the specific (e.g., comparing the environmental performances of several possible designs for a MSW management system in a particular community). There are some, such as Eriksson et al. (2005), which focus on identifying the option with the lowest environmental, energy and economic costs. Others, such as Solano et al. (2002a,b) and Carlsson Reich (2005), use case studies to examine the utility of LCA models and reveal both the possibilities and the limitations of linking economic information to LCAs of MSW. Emery et al. (2007) compare scenarios that would fulfill waste management regulations, including the European Union Landfill Directive. Finnveden et al. (2005)/Moberg et al. (2005) employ a case study to evaluate whether or not the LCA results are compatible with the ‘hierarchy of waste.’ The papers limited their ‘waste hierarchy’ evaluation to recycling, incineration and landfilling, and do not include scenarios addressing activities to reduce waste generation or to reuse waste materials.

2.4.2.1 Comparisons of MSW management systems

All of the studies are comparative LCAs that evaluate the environmental emissions and/or performance of various types of solid waste management systems. Some compare one scenario with a defined baseline, often landfilling, while most compare a number of options to one another. Most of the reviewed LCAs include the landfilling, recycling, and thermal treatment of MSW, with fewer studies modelling composting and anaerobic digestion. Source reduction is omitted from all but one (Rodriguez-Iglesias et al. 2003) of the studies reviewed. 17

Only one of the LCAs (Connsonni et al. 2005a, b) evaluates scenarios to quantify the effect of differences in waste treatment efficiencies relative to the scale of treatment. This LCA claims that the economic cost and efficiency of an incinerator varies with the amount of waste produced as well as on the size of the facility. Similarly, a landfill may need to reach a particular size in order to extract biogas.

2.4.3 Functional units

The functional unit is fundamental to the understanding of the results of an LCA, and provides a common basis for the comparison of results (Consonni et al. 2005b; Rebitzer et al. 2004). It exists as a reference unit to which the input and output data are normalized (ISO 2006). For LCAs of MSW, it ensures that all of the environmental emissions are based on identical inputs to each waste management system. Functional units are also associated with the usable products generated during MSW management, including electricity, heat and compost. This component of the study addresses the quantity of reviewed LCAs that provide an explicit definition of the primary (i.e., non-compensatory) functional unit. This explicit definition requires a mention of the term ‘functional unit,’ along with its description, for the MSW management system studied through the use of LCA. The ISO (2006) has deemed that “the functional unit shall be clearly defined and measurable.” Only 11/20 of the reviewed studies provide an explicit definition of the primary functional unit of the LCA, perhaps indicating that many of the LCA practitioners felt that the primary functional unit used was self-evident (Fig. 1). Although the primary functional units defined in the reviewed LCAs of MSW are variants on ‘tonnes of MSW treated per year,’ Rodriguez-Iglesias et al. (2003) display their results relative to the volume of MSW treated. 18

Yes No Not Mentioned / Unclear

Explicit definition of primary functional unit

Inclusion of life cycle emissions from the production of capital / infrastructure

Inclusion of emissions from MSW transport

0 2 4 6 8 101214161820 Number of Reviewed LCAs of MSW

Figure 1 Primary functional units, the inclusion of life cycle emissions from the production of capital / infrastructure, and the inclusion of emissions from MSW transport in the reviewed LCAs of MSW management systems. Additional detail is available in Appendix 2.

2.4.4 System boundaries

The LCA system boundary is the interface between the product or waste management system and the environment or other product systems, determining which unit processes are included within the LCA. The system boundaries must account for time, space and function (Eriksson et al. 2002). The ISO 14044 standard, when addressing the goal and scope definition component of LCA, states that “any decisions to omit life cycle stages, processes, inputs or outputs shall be clearly stated and the reasons and implications for their omission shall be explained” (ISO 2006). The life cycle stages of MSW management commonly include: (1) collection; (2) transportation to a sorting facility; (3) sorting; (4) transportation to a treatment facility, and (5) treatment - potentially including recycling, biological treatment, thermal treatment and landfilling. Unusual cut-off criteria for system boundaries are observed in Tan and Khoo (2006) and Schmelev and Powell (2006). Tan and Khoo (2006) have the only LCA 19

which omits the net environmental benefits from the processing of recyclable wastes sent overseas, since the authors only include the management of MSW within the geographic boundaries of Singapore. Schmelev and Powell (2006) take into account the spatial dimension of the environmental emissions and ecological sensitivity of the affected areas. The authors consider LCA as one component of ecological–economic modelling, combining LCA with an environmental impact assessment and a geographic information system framework (Schmelev and Powell 2006).

2.4.4.1 The ‘cradle’ and ‘grave’ of waste

LCA practitioners tend to view the ‘cradle’ in the life cycle of MSW as the moment that an item becomes, or is perceived as, valueless and is thrown out, sent for recycling or for biological treatment. The ‘grave’ of waste is reached when its value is restored as a useable material or energy, when it is transformed into air and water emissions (e.g., CO2 and leachate), or into inert landfilled material (Özeler et al. 2006). For recycled waste materials, LCAs of MSW can also claim to take a ‘cradle to cradle’ approach in defining the waste management system. Most of the reviewed studies do not provide clear definitions of the cradle and grave of the life cycle of the MSW management system under study, although they are generally evident through the calculations undertaken in each LCA. Only one LCA (Rodriguez-Iglesias et al. 2003) compared MSW management scenarios that have different total quantities of MSW collected from the same population, suggesting a ‘cradle’ in advance of waste generation, possibly resulting from the implementation of waste prevention and product reuse activities. A disparity in the depiction of the ‘grave’ is particularly evident for the landfilling of MSW. Many of the LCAs assume that the waste would completely decompose (e.g., Aye and Widjaya 2006) or become inert 100 years after being landfilled (e.g., Buttol et al. 2007), and would no longer contribute to leachate and biogas emissions. Only one reviewed LCA (Finnveden et al. 2005/Moberg et al. 2005) addresses this landfilling issue in its sensitivity analysis (non-explicit - see Section 2.4.7).

2.4.4.2 Life cycle environmental emissions from the production of capital / infrastructure

For their operation, waste management systems require inputs of energy and materials, as well as capital equipment and infrastructure such as collection vehicles and thermal treatment facilities. All of these inputs have environmental emissions associated with their own product life cycles. Many of these emissions, also known as secondary environmental burdens, tend to be excluded from LCAs of MSW since they are assumed to be relatively small in comparison to primary burdens (McDougall et al. 2001). Finnveden et al. (2005) and McDougall et al. (2001) claim a ‘rule of thumb’ suggesting that secondary environmental burdens comprise up to 10% of the life cycle impacts of MSW management. In some cases, these burdens could be highly significant to the overall results of the LCA. For example, the results from Consonni et al. (2005b) show that excluding the greenhouse gas emissions resulting from the construction of a large incinerator (amortized over the expected life span of the facility) would increase the life cycle global warming potential (GWP) benefit of MSW incineration scenarios by between 50% and 300%. However, the estimated GWP from the construction of the incinerator is less than 10% of the gross emissions from incinerating the MSW (Consonni et al. 2005b). Only three of the reviewed LCAs (Buttol et al. 2007; Consonni et al. 2005a, b; Mendes et al. 2004) indicate explicitly that the life cycle environmental emissions from the production of capital equipment and infrastructure are included within the system boundaries of their studies (Fig. 1). More than half of the LCAs (11/20) either do not mention whether or not these emissions are within the system boundaries, or are unclear about this issue. The remainder (6/20) exclude these emissions.

2.4.4.3 Environmental emissions from MSW transportation

Three of the reviewed articles omit the environmental emissions from the transportation of MSW (Fig. 1). Hong et al. (2006) state that this omission is justified because the environmental emissions from transportation would not differ between the 21

waste management scenarios. Mendes et al. (2004) claim that the environmental emissions from transportation would be negligible relative to those from the waste treatment components, citing experiences from other LCAs of MSW. In contrast, Chaya and Gheewala (2007) do not provide a reason for omitting transportation. The results from Consonni et al. (2005b), the only reviewed LCA that provides estimates of the impacts from both the transportation component of the LCA and the production of capital equipment, are of particular interest. The study estimates that the GWP resulting from the transportation of MSW to an incinerator is much less than the GWP from the construction of the incinerator (amortized over the expected life span of the facility). Nevertheless, the results presented in Section 4.4.2 of this comparative analysis show that the life cycle environmental burdens of capital equipment are much more likely to be omitted than the impacts from the transportation component of the LCA. Some publications, such as Finnveden and Ekvall (1998), have claimed that the environmental emissions from the transportation component of LCAs of MSW do not affect the overall results so long as the transportation systems are relatively efficient. Since Beigl and Salhofer (2004), for example, find the transportation component to be significant in determining which MSW management scenario generates greater ecological benefits, LCA practitioners should perhaps be wary of removing the transportation stage from within the system boundary. However, excluding this component from scenario comparisons may be justifiable if it is equivalent, or close to equivalent, in all of the scenarios.

2.4.4.4 Selection of energy sources

MSW management systems consume electricity and fossil fuels, and may generate usable heat and electricity from thermal treatment systems, as well as from the combustion of biogas collected from landfills and anaerobic digestion facilities. The assumptions relating to the method of generating the energy both consumed and displaced by MSW management systems may have a substantial effect on the results of the LCA (Björklund and Finnveden 2005). 22

Eight of the 20 LCAs either do not mention, or are unclear about whether or not the environmental emissions from electricity consumption and displacement are based upon the average or the marginal source of electricity. For the remaining 12 studies, six of them use electricity marginals in which the electricity produced from waste treatment systems replaces electricity produced by the ‘least-efficient’ means of production (e.g., coal). The other six studies base the environmental emissions upon the average mix of electricity sources in the region of interest. None of the LCAs conduct a sensitivity analysis of these assumptions. Heat displacement is mentioned to a lesser degree than electricity in the selected LCAs, possibly because it was uncommon to have a facility located nearby a thermal treatment facility that could use the generated heat. Consonni et al. (2005a, b) find that the use of the heat co-product has a substantial effect on the life cycle environmental emissions of the waste management system. Half of the reviewed LCAs use environmental emission estimates for electricity grids located in the same country as the MSW management systems. Eight of the remainder do not mention which electricity network is selected, while one chooses the marginal source for the entire European Union. One LCA (Beigl and Salhofer 2004) uses data for an electricity network in a different country from the MSW management system it examines.

2.4.5 Environmental impacts and LCIA

The life cycle impact assessment (LCIA) encompasses the creation of impact categories, the assignment of inventory data to specific impact categories (classification), and the modelling of the magnitudes of the inventory data within impact categories (characterisation) (Bengtsson and Steen 2000). LCIA may also include the valuation or single score weighting of impacts in order to estimate the environmental performance of a particular scenario. Although LCIA is more useful to policy-makers than is the life cycle inventory (Craighill and Powell 1996), it is less objective and more difficult to undertake. Only one of the reviewed LCAs (Solano et al. 2002a, b) confines the analysis to the inventory stage. 23

All of the LCIAs provide quantitative estimates of the environmental impacts of the waste management scenarios. This necessitates the sorting and aggregation of environmental loadings (e.g., carbon dioxide and methane emissions), identified in the inventory stage, into the impact categories (e.g., GWP).

2.4.5.1 Impact categories

The choice of impact category results to display is subjective, although the most common in the reviewed articles include global warming potential, acidification potential, eutrophication of surface water and resource consumption (Fig. 2). Toxicity impacts are much less popular, with 8/20 addressing human toxicity and 6/20 including ecotoxicity. The numerous impact category omissions may not have permitted several of the reviewed LCAs to “reflect a comprehensive set of environmental issues,” as is required in the ISO 14042 standard for life cycle impact assessment (ISO 2000).

2.4.5.2 Characterization of impacts

Analysis of impact magnitudes in each category takes place during the characterization stage, which applies exclusively to LCIAs. The typical methods of characterizing the composition of each impact include Eco-Indicator 95 and 99, World Reserves Life Index, the Danish EDIP method and the Dutch model USES-LCA. Only one of the studies (Schmelev and Powell 2006) appears to take local ecosystem sensitivity into account when characterizing each environmental impact, and few mention it at all. Moreover, the choice of characterization method is almost always left unjustified. 24

these impacts 6

2 Number of reviewed LCAs including LCAs including Number of reviewed 0 AP GWP Eut RC HT PO Etx SO HM Impact category

Figure 2 Impact categories included in the reviewed LCAs of MSW management systems. Acronyms: AP-Acidification potential; GWP-Global warming potential; Eut-Eutrophication; RC-Resource consumption; HT-Human toxicity; PO-Photochemical oxidants or ozone; Etx-Ecotoxicity; SO- Stratospheric ozone; and HM-Heavy metals. Additional detail is available in Appendix 2.

2.4.5.3 Single score weighted valuations of impacts

Single score weighted valuations of each impact category may take place in order to meet the study objectives, especially those referring to the minimization of environmental impacts or costs. The weighting of environmental impacts in homogeneous units allows a quantitative comparison of the impacts. For example, weighting permits the LCA practitioner to evaluate whether or not the climate change effect of a particular waste treatment scenario is greater than its eutrophication potential. One can also aggregate all of the environmental impacts of each scenario in order to compare their overall environmental performances. ISO (2006) advises LCA practitioners to acknowledge the lack of a scientific basis for the application of weighting techniques which reduce LCA results to a single overall score. In addition, ISO (2006) claims that weighting should not be used in LCAs that make comparative assertions intended to be disclosed to the public. Six of the reviewed LCAs perform weightings across impact categories (Fig. 3). Typically, an economic evaluation of environmental impacts is associated with these weightings (e.g., Carlsson Reich 2005; Finnveden et al. 2005; Morris 2005). Tan and Khoo (2006) do not specify the weighting method used, only that it is available in the SimaPro LCA software package. Carlsson Reich (2005) compares the results from three 25

weighting methods (ECON95, WTPGPS 2000, and EcoTax99), all of which use an economic framework. This paper indicates that the results using these three methods show substantial differences in absolute cost estimates, while the relative costs associated with the environmental aspects studied are similar. Of the five LCAs that generated weighted impacts for more than one scenario, four are in agreement with order of the waste treatment components of the ‘hierarchy of waste.’ This hierarchy includes waste prevention, product reuse, recycling, biological treatment, thermal treatment and landfilling (McDougall et al. 2001). The location of each of these forms of waste management in the hierarchy is dependent upon the capacity of each to generate less waste, conserve additional raw materials, and reduce energy use and environmental impacts (Maclaren 2004). Finnveden et al. (2005) and Tan and Khoo (2006) estimate that recycling and biological treatments of MSW have better environmental performances than thermal treatments, and that landfilling has the worst performance of all. Hong et al. (2006) find similar results, although recycling is included in each scenario. Carlsson Reich (2005), who compares the results from three weighting methods, finds that landfilling shows the worst environmental performance. Of the biological treatments evaluated, Carlsson Reich (2005) estimates that anaerobic digestion is superior to composting. Moreover, he observes a lack of consistency in the comparative performances of the incineration, biological treatment and recycling scenarios, except for the scenario that includes the recycling of plastics, which always shows the best environmental performance. The LCA by Morris (2005) does not provide the overall environmental impact results for more than one scenario. In contrast to the results from the other LCAs with single score impact weightings, Rodriguez-Iglesias et al. (2003) determine that the incineration of MSW with energy recovery is approximately 12 times worse than the other weighted scenarios. This finding is primarily due to the high normalization factors assigned to the emission of heavy metals and carcinogens. 26

Yes No

Inclusion of a single score weighted valuation of impacts

Inclusion of an explicit sensitivity analysis

0 2 4 6 8 10 12 14 16 18 20 Number of Reviewed LCAs of MSW

Figure 3 The inclusion of single score weighted valuations of impacts and explicit sensitivity analyses in the reviewed LCAs of MSW management systems. Additional detail is available in Appendix 2.

2.4.6 Types and sources of data

All of the reviewed LCAs use some data from the scientific literature and/or from generic emission coefficient databases to determine environmental emissions when direct measurements are not available. Much of the data collected for each case study are related to waste material generation and their flows, and full lists of the input data for each LCA are not provided. In many instances, peer review of the results would not be possible without obtaining additional data through personal communication with the author(s) of the LCAs. In the reviewed LCAs, it is much more common to use emission databases to calculate the environmental emissions from waste management than to use direct measurements from the specific waste treatment plants under consideration. Chaya and Gheewala (2007) provide direct emission measurements from an incinerator. Arena et al. (2003) use some measured emission data from landfills, design data from incinerators not yet in operation, and data from incinerators located in other regions of Italy. 27

Geographic differences between the source of data and the location of the study are common. For example, Aye and Widjaya (2006) select an Australian emissions database for their LCA that addresses Indonesia. Some authors, such as Di Maria and Fantozzi (2004), do not mention the sources of data used.

2.4.7 Sensitivity analysis

A sensitivity analysis may be undertaken to determine the effects on the results of altering selected model or assessment parameters (i.e., quantifiable data components, such as waste generation rate and composition of waste). It consists of varying the value of a model parameter across its range of technically reasonable values. Sensitivity analyses, explicitly defined (i.e., the expression ‘sensitivity analysis’ is used), are pursued in 4/20 of the LCAs reviewed (Fig. 3). Eriksson et al. (2005) refer to a sensitivity analysis, but do not display results from this analysis in the article. Although Arena et al. (2003), Carlsson Reich (2005), Di Maria and Fantozzi (2004), Emery et al. (2007), and Rodriguez-Iglesias et al. (2003) do not use the expression ‘sensitivity analysis,’ their evaluations of the effects of modifying various parameters of waste management scenarios on the results can be considered sensitivity analyses, and thus more than doubles the sensitivity analyses to 9/20. Finnveden et al. (2005) claim a broader depiction of the sensitivity analysis: “… extensive use of scenarios can be regarded as a sensitivity analysis of the studied system.” This broader depiction would further increase the number of reviewed LCAs incorporating a sensitivity analysis. At a minimum, Consonni et al. (2005), Finnveden et al. (2005)/Moberg et al. (2005), Özeler et al. (2006), Solano et al. (2002), as well as Eriksson et al. (2005), which possesses scenarios that vary the proportions of MSW allocated to the various waste treatments, would be added to the list. The model parameters targeted by sensitivity analyses include MSW transportation distances, sorting efficiency, waste treatment capacity, impact weighting methods, recycling program participation, recovery rates of recyclable materials, waste composition, project time horizons, and costs of MSW transportation and recycling.

2.4.8 Economic costs of MSW treatment

An economic analysis with similar system boundaries to an LCA is known as a financial ‘life cycle costing’ (LCC) (Carlsson Reich 2005). This type of analysis views the waste management system as a single economic actor and facilitates the estimation of the annual life cycle economic costs of waste treatment systems irrespective of who bears the financial cost. It allows the financial impact to be listed along with the measured or estimated environmental impacts (Carlsson Reich 2005). Although the International Organization of Standardization (2006) suggests that LCAs do not typically include an economic component, a substantial number of the reviewed studies contain exactly that. Just under half of the LCAs of waste (8/20) incorporate financial LCCs which estimate the financial cost of the waste treatment activities (Fig. 4). The costs tend to focus on the processes directly associated with the treatment of MSW, and exclude indirect costs, such as the effect on nearby property values of constructing a landfill, incinerator or composting facility. System boundary inconsistencies are evident in some of those studies encompassing both an LCA and an LCC. Of particular note is the system boundary inconsistency relative to the environmental and economic costs of producing capital equipment. Half of those LCAs with financial LCCs include these costs (Fig. 4). However, some studies (Aye and Widjaya 2006; Shmelev and Powell 2006) exclude the environmental impacts from the production of capital equipment, yet incorporate their economic costs. External (non-market) environmental costs and benefits are not taken into account in standard financial assessments. One quarter of the studies (5/20) undertake an environmental LCC, which attributes economic values to the environmental impacts of each waste treatment scenario (Fig. 4). Eriksson et al. (2005) use a willingness-to pay method, while Morris (2005) uses the average value for pollutant emissions, derived from a review of four studies listing the cost estimates of market trades for pollutant emissions permits or agreements. Three LCAs have a combination of a financial LCC and an environmental LCC, which would likely possess greater relevance for waste managers 29 than the simple monetary valuation of the life cycle environmental impacts of each scenario.

Yes Yes (including capital equipment) Yes (excluding capital equipment) No

Economic costs of MSW management activities

Economic valuation of environmental impacts of MSW management

0 2 4 6 8 10 12 14 16 18 20 Number of Reviewed LCAs of MSW

Figure 4 The inclusion of the economic costs of MSW management activities and the economic values of the environmental impacts of MSW management in the reviewed LCAs of MSW management systems. Additional detail is available in Appendix 2.

2.5 Comparison of results

In light of the waste hierarchy’s generalized depiction of the preferential order of waste treatment techniques, the results from different LCAs were compared statistically to illustrate the differences between the results of landfilling, thermal treatment and mixed treatment LCA scenarios. Out of the 20 reviewed LCAs, 8 provide data on acidification potential (AP) in SO2 equivalents, 11 supply global warming potential (GWP) estimates, and the same number have results for net energy use (NEU), also designated as ‘resource consumption.’ Statistical indicators are used to compare the NEU, GWP and AP results of each type of MSW management scenario and are illustrated in Figs. 5, 6 and 7. These include (1) the median, which provides the typical value; (2) the interquartile range (IQR), in which 50% of all of the cases have values within this range (the 25th to the 75th percentile); (3) the “whiskers,” or the vertical lines drawn from the edges of the IQR boxes which indicate the largest and smallest values within 1.5 times the IQR length from the upper and lower boundaries of the IQR; and (4) 30 the outlier values, which are within 1.5 and 3 IQR lengths from the upper and lower boundaries of the IQR. The extreme values, which are greater than 3 IQR lengths from the upper and lower boundaries of the IQR, are not shown in the figures. Some of the excluded studies do not publish the required data, while the others do not publish it in a format that permits a comparison of the results on a per tonne (MSW) basis. One scenario from Morris (2005) is excluded from the comparative analysis of results because it addresses only recyclable materials out of all of the MSW streams. Table 2 lists and describes the types of MSW management scenarios in the reviewed LCAs for which there are comparable AP, GWP and NEU data.

Table 2 Acidification potential (AP), global warming potential (GWP), and net energy use (NEU) values for each MSW treatment scenario. Acronyms AD – anaerobic digestion; BMT – biological and mechanical pre-treatment; B – biological treatment; BW – biodegradable waste; Inc. – incineration; IER – incineration with energy recovery; L – landfilling; LER – landfilling with energy recovery; LP – large plant; M – mixed treatment; Pref. order – preferential order of treatment scenario types in the LCA; R – recycling; RDF – refuse derived fuel; SLOF – stabilization and landfilling of organic fraction; SP – small plant; T – thermal treatment; UW – untreated waste. Note Only those LCAs of MSW providing comparable data on NEU, GWP and AP are listed in Table 2. A small minority of the listed values were derived through the interpretation of graphs.

Source AP GWP NEU Type of Assigned number of treatment scenario and its description (kg SO2 (kg CO2 (GJ / t treatment eq. / t eq. / t MSW) scenario MSW) MSW) Arena et -0.44 490 -0.67 L (1.1) LER al. 2003 -3.66 95 -4.95 T (1.2) Inc. of RDF, with energy recovery -4.60 46 -6.35 T (1.3) IER of unsorted and untreated waste Pref. order T, L T, L T,L Aye and 0.03 - - L (2.1) open dumping Widjaya -0.11 M(BL) (2.2) composting in labour intensive plants 2006 -0.10 M(BL) (2.3) composting in a centralized plant -0.50 M(BL) (2.4) AD -0.51 L (2.5) LER Pref. order L, M(BL), L Beigl and -0.55 152 -1.66 M(BLR) (3.1) R - bring system Salhofer -0.68 141 -1.67 M(BLR) (3.2) R - curbside collection 2004 -0.16 145 -1.48 M(BLR) (3.3) non-recycling – mechanical biological Pref. order N/A N/A N/A Buttol et - -28 -3.62 M(BLRT) (4.1) business as usual al. 2007 -52 -4.15 M(BLRT) (4.2) additional incinerator in “area 2” of Bologna district -43 -3.98 M(BLRT) (4.3) additional incinerator in “area 5” of Bologna district Note: All scenarios include IER, LER, composting and R. Pref. order N/A N/A N/A Chaya and 2.37 - -0.56 T (6.1) IER Gheewala 1.57 -3.58 M(BL) (6.2) AD (63%) and L (37%) 2007 Pref. order M(BL), T N/A M(BL), T Consonni -2.14 101 -5.61 T (7.1) SP – Inc. of UW et al. 2005 -1.92 111 -5.02 T (7.2) SP – Inc. of pre-treated waste, SLOF [a, b] -1.38 122 -3.54 T (7.3) SP – Inc. of RDF -1.78 102 -4.46 T (7.4) SP – Inc. of RDF, SLOF -3.22 -53.7 -7.72 T (7.5) LP – Inc. of UW (steam for electricity generation) 31

-2.91 -134 -8.99 T (7.6) LP – Inc. of UW (30% of steam for district heating) -2.59 -212 -10.23 T (7.7) LP – Inc. of UW (60% of steam for district heating) -2.94 -35.8 -7.01 T (7.8) LP – Inc. of pre-treated waste, SLOF -2.27 -5.8 -5.28 T (7.9) LP – Inc. of RDF -2.69 -29.3 -6.23 T (7.10) LP – Inc. of RDF, SLOF Pref. order N/A N/A N/A Eriksson et 1.17 339 - T (10.1) IER al. 2005 1.14 355 T (10.2) 90% IER; 10% LER 1.79 349 M(BT) (10.3) Inc. and AD of 70% of BW – biogas for fuelling buses 2.16 355 M(BT) (10.4) Inc. and AD of 70% of BW – biogas for heat and power 1.48 347 M(BT) (10.5) Inc. and AD of 70% of BW – biogas for fuelling cars 1.09 233 M(RT) (10.6) Inc. and R (70%-80% of HDPE and LDPE – excl. residential LDPE) 1.13 351 M (RT) (10.7) Inc. and R (70%-80% of cardboard) 0.99 508 L (10.8) LER Pref. order L, M(RT), M(RT), T, N/A T, M(BT) M(BT), M(RT), [T, M(BT)] Finnveden - -500 -32.9 M(BR) (11.1) R and AD et al. 2005 500 -17.9 T (11.2) IER / Moberg 1900 -2.99 L (11.3) LER et al. 2005 Pref. order N/A M(BR), M(BR), T, L T, L Hong et al. 0.14 461 0.08 L (12.1) L (no mention of energy recovery) 2006 0.51 229 -0.32 T (12.2) IER 0.00 13.9 0.12 M(BR) (12.3) BMT – compost 0.32 151 M(RT) (12.4) BMT – incineration 0.07 315 M(LR) (12.5) BMT – landfill (no mention of energy recovery) Pref. order M(BR), M(BR), T, L, M(LR), L, M(RT), T, M(BR) M(RT), T M(LR), L Kirkeby et - -732 -31.9 M(RT) (13.1) Inc., AD (<10%) and R (81 000 tonnes MSW) al. 2006 -740 -32.1 M(RT) (13.2) Inc. and R (81 000 tonnes MSW) Pref. order N/A N/A N/A Mendes et 0.13 600 -3.34 T (14.1) Inc. with ash disposal al. 2004 0.20 625 -2.84 T (14.2) Inc. with ash melting 0.20 650 -2.06 T (14.3) Inc. with ash reuse system – brick production 0.40 920 2.43 L (14.4) L (no energy recovery) 0.30 900 1.83 L (14.5) LER Pref. order T, L T, L T, L Morris - -50 -1.0 L (15.1) LER [SLO case study] 2005 Pref. order N/A N/A N/A Solano et - 19 -1.79 M(LR) (19.1) minimum cost al. 2002 [a, -130 -10.4 T (19.2) minimum GHG emissions b] Pref. order N/A T, M(LR) T, M(LR)

Each MSW management result is classified into one of three types of treatment scenarios: landfilling, thermal treatment, and mixed treatment. For the landfilling scenarios, it is assumed that all of the MSW is deposited in a landfill, either with or without energy recovery. The thermal treatment scenarios comprise those in which almost all of the MSW is treated thermally, generally greater than 90% by mass. Those scenarios designated as ‘mixed’ include at least two of the following in non-negligible 32 amounts (i.e., greater than 10% by mass of the MSW collected): thermal treatment, landfilling, recycling and/or biological treatment. Although it would have been preferable to isolate into separate categories the scenarios that adopt primarily biological treatments and recycling, it is not practical because: (1) these treatments apply to different MSW streams (i.e., biological treatment of organic materials; recycling of recyclable materials); and (2) classifying the mixed scenarios into even smaller categories that specify the proportions of MSW allocated to each specific treatment type would reduce the average number of scenarios to only three per category. In Table 2, those forms of MSW treatment which apply to more than 10% (by mass) of the MSW in each mixed treatment scenario are listed in brackets under the ‘type of treatment scenario’ column. However, all mixed treatments scenarios are incorporated into one category for the statistical analysis. In addition to the statistical comparison of the results for all of the reviewed LCAs, the preferential order (i.e., lowest environmental impacts to highest) of the MSW treatment scenario types within each LCA is also defined for the AP, GWP and NEU impact indicators (Table 2). Those LCAs in which the following scenario types show increasing net impacts: recycling, biological treatment, thermal treatment and/or landfilling; possess results in agreement with the waste hierarchy. Nevertheless, the results of this comparison are not intended to be used to prescribe specific waste treatments for particular waste management systems as each system has its own characteristics in terms of technologies employed, waste composition and distances to treatment facilities. These variances have the potential to affect the net impacts of each type of treatment relative to one another.

2.5.1 Acidification potential

The results from the 42 scenarios with acidification figures range from −4.60 to

2.37 kg of SO2 equivalents per tonne of MSW treated. AP values are available for 7 landfilling scenarios from 5 LCAs, 15 mixed treatment scenarios from 6 LCAs and 20 thermal treatment scenarios from 6 LCAs. Only the thermal treatment scenarios indicate a median reduction in acidification potential (−1.85), while the landfilling scenarios have 33

the highest median acidifying emissions (0.14) (Fig. 5). The MSW treatment scenarios with the greatest variability and ranges of values are in the thermal treatment category, while the results from the landfilling category show the least. When examining the preferential order (in terms of environmental performance) of the waste treatment scenario types within each study, only the AP results from 2 out of 6 of the reviewed LCAs are in agreement with the waste hierarchy.

Only those AP figures provided in a comparable format (i.e., SO2 equivalents) are incorporated into this comparative analysis. AP estimates from one of the LCAs (Özeler et al. 2006) are excluded since they are an order of magnitude or two greater than all of the other results, and can perhaps be considered errors.

Figure 5 The median values, interquartile ranges, whiskers and outliers of the acidification potential values for MSW management systems classified into landfilling, mixed treatment and thermal treatment scenarios.

2.5.2 Global warming potential

The GWPs of the scenarios from the reviewed LCAs range from −740 to 1900 kg

CO2 equivalents emitted per tonne of MSW treated. GWP values are available for 7 landfilling scenarios from 6 LCAs, 18 mixed treatment scenarios from 8 LCAs and 20 34

thermal treatment scenarios from 7 LCAs. By far the highest median GWP is associated with landfilling scenarios (508), with thermal treatment scenarios showing the lowest (101) (Fig. 6). The GWP IQR for landfilling is higher than for the other types of scenarios, and does not overlap with them. In 3 out of 6 of the reviewed LCAs, the preferential order of the waste treatment scenario types are in agreement with the waste hierarchy. In order to be consistent, only those GWP estimates based upon a 100 year time horizon are used. Some LCAs, including Emery et al. (2007) and Aye and Widjaya (2006), use 20 and 50 year time horizons, respectively, to depict the GWP data, and therefore, these values could not be included in the comparison.

Figure 6 The median values, interquartile ranges, whiskers and outliers of the global warming potential values for MSW management systems classified into landfilling, mixed treatment and thermal treatment scenarios.

2.5.3 Net energy use

The NEU results from the selected LCAs range from −32.9 to 2.43 GJ per tonne of MSW treated. There are NEU values for 6 landfilling scenarios from 5 LCAs, 12 35

mixed treatment scenarios from 7 LCAs and 18 thermal treatment scenarios from 7 LCAs. All of the median values for each type of scenario indicate a net energy gain, with the largest net gain for thermal treatment scenarios (−5.28 GJ/tonne MSW) and the smallest for landfilling (−0.30) (Fig. 7). Not surprisingly, the results from the mixed treatment scenarios indicate the greatest variability, more than twice the variability of the landfilling and thermal treatment categories. This is due, in part, to the diversity of MSW management scenarios within this category. Similar to the GWP statistical results, the findings for NEU are in agreement with the waste hierarchy inasmuch as a thermal treatment scenario appears to be an improvement over landfilling. For the comparison of results within each LCA, the findings are less convincing. In half of the reviewed LCAs, the preferential order of the waste treatment scenario types are in agreement with the waste hierarchy. The results from the mixed treatment scenarios in Kirkeby et al. (2006) and Finnveden et al. (2005)/Moberg et al. (2005) show a much larger net energy gain than those of the other LCAs, with a mean NEU (−32.3) more than 13 times the mean NEU (−2.42) when excluding the three scenarios. This discrepancy can be explained by the higher levels of recycling in the scenarios from these two LCAs than the levels in most of the others.

Figure 7 The median values, interquartile ranges, whiskers and outliers of the net energy use values for MSW management systems classified into landfilling, mixed treatment and thermal treatment scenarios.

2.6 Discussion

The selected LCAs address MSW management systems at mostly local and regional scales in a wide variety of locations. Very few of the LCAs are consistent in following the general requirements and guidelines for LCA stipulated by the ISO. The ISO (2006) 14044 LCA international standard is even more stringent for studies that are intended to be used in comparative assertions. All of the reviewed LCAs make comparative assertions, which according to ISO (2006), necessitates a sensitivity analysis and a comprehensive set of impact category indicators, but precludes the weighting of environmental impacts. Perhaps of greater concern is the lack of clarity in identifying the system boundaries of the LCAs, especially regarding the inclusion or exclusion of the environmental burdens associated with the production of capital/infrastructure. As Finnveden et al. (2005) claim, the uncertainty associated with choices relating to LCA methodology tends to be larger than that associated with data and precision. 37

The display of results from the AP and GWP impact categories is more common than from any other category types. The lower levels of use of toxicological impact categories is perhaps not surprising since the uncertainties associated with them are large relative to the more popular categories such as GWP (Moberg et al. 2005). Only four types of impacts are shared by more than half of the reviewed LCAs. The LCA practitioners may not have selected a “sufficiently comprehensive set of category indicators,” as is specified by ISO (2006) for LCIAs intended to be used in comparative assertions for public disclosure. The very substantial variation in the application of LCA methodology, system boundaries and assumptions, could lead one to expect no consensus relating to the environmental performance of MSW treatments in mixed-waste systems. However, the single score weighted results of the reviewed LCAs tend to confirm the validity of the ‘hierarchy of waste.’ Similarly, the results from the statistical analysis of NEU, AP and GWP indicate that thermal treatment scenarios are inclined to have a better environmental performance than landfilling, while the results for the more loosely defined mixed treatment scenarios are less clear. The GWP values demonstrate the smallest differences in the relative variability (IQR) between the three scenario categories. The NEU and AP results from landfilling scenarios also show a much lower variability than the results associated with the other scenario categories. The high variation in the waste treatment compositions of the mixed treatment scenarios is evident in the results. In contrast, the more homogeneous composition of the landfilling and thermal treatment scenarios mostly results in less statistical variation within these categories. One can therefore have more confidence in the results comparing the landfilling and thermal treatment scenario impacts because their treatment profiles are less wide-ranging.

2.7 Conclusion

The results from this comparative analysis suggest that many LCA practitioners do not make clear the decisions made during the goal and scope definition phase of their LCAs of MSW management. Increasing the number of impacts evaluated should be a 38

priority. Meeting the ISO 14044 requirements and guidelines for LCA would provide the readers of published LCAs with a much more informed understanding of the methodologies, system boundaries and assumptions of the studies, and make it easier to interpret results based on comparisons of different LCAs. An analysis of the quantitative effect of particular system boundary and other methodological choices on the results of the reviewed LCAs was necessarily outside the scope of this paper, since the magnitude of the effect on the overall results would differ with each MSW management system studied. However, those choices made by the LCA practitioners do not appear to influence significantly the order of the preferred options for MSW treatment.

2.8 References

Arena U, Mastellone M, Perugini F. 2003. The environmental performance of alternative solid waste management options: a life cycle assessment study. Chem Eng J 96:207–22.

Aye L, Widjaya E. 2006. Environmental and economic analyses of waste disposal options for traditional markets in Indonesia. Waste Manage 26(10):1180–91.

Beigl P, Salhofer S. 2004. Comparison of ecological effects and costs of communal waste management systems. Resour Conserv Recycl 41:83-102.

Bengtsson M, Steen B. 2000. Weighting in LCA — approaches and applications. Environ Prog 19(2):101–9.

Björklund A, Finnveden G. 2005. Recycling revisited—life cycle comparisons of global warming impact and total energy use of waste management strategies. Resour Conserv Recycl 44(2):309–17.

Buttol P, Masoni P, Bonoli A, Goldoni S, Belladonna V, Cavazzuti C. 2007. LCA of integrated MSW management systems: case study of the Bologna District. Waste Manage 27(8):1059–70.

Carlsson Reich M. 2005. Economic assessment of municipal waste management systems — case studies using a combination of life cycle assessment (LCA) and life cycle costing (LCC). J Clean Prod 13:253–63.

Chaya W, Gheewala SH. 2007. Life cycle assessment of MSW-to-energy schemes in Thailand. J Clean Prod 15(15):1463–8.

Consonni S, Giugliano M, Grosso M. 2005a. Alternative strategies for energy recovery from municipal solid waste. Part A: mass and energy balances. Waste Manage 25(2): 123–35.

Consonni S, Giugliano M, Grosso M. 2005b. Alternative strategies for energy recovery from municipal solid waste. Part B: emission and cost estimates. Waste Manage 25(2): 137–48.

Craighill AL, Powell JC. 1996. Lifecycle assessment and economic evaluation of recycling: a case study. Resour Conserv Recycl 17:75–96.

Denison RA. 1996. Environmental life-cycle comparisons of recycling, landfilling and incineration: a review of recent studies. Annu Rev Energy Env 21:191–237.

Di Maria F, Fantozzi F. 2004. Life cycle assessment of waste to energy micro-pyrolysis system: case study for an Italian town. Int J Energy Res 28:449–61.

Emery A, Davies A, Griffiths A, Williams K. 2007. Environmental and economic modelling: a case study of municipal solid waste management scenarios in Wales. Resour Conserv Recycl 49:244–63.

Eriksson O, Carlsson Reich M, Frostell B, Björklund A, Assefa G, Sundqvist J-O, et al. 2005. Municipal solid waste management from a systems perspective. J Clean Prod 13:241–52.

Eriksson O, Frostell B, Björklund A, Assefa G, Sundqvist J-O, Granath J, et al. 2002. ORWARE — a simulation tool for waste management. Resour Conserv Recycl 36(4):287–307.

Finnveden G, Ekvall T. 1998. Life-cycle assessment as a decision-support tool—the case of recycling versus incineration of paper. Resour Conserv Recycl 24:235–56.

Finnveden G, Johansson J, Lind P, Moberg A. 2005. Life cycle assessment of energy from solid waste — part 1: general methodology and results. J Clean Prod 13:213–29.

Hong RJ,Wang GF, Guo RZ, Cheng X, Liu Q, Zhang PJ, et al. 2006. Life cycle assessment of BMT based integrated municipal solid waste management: case study in Pudong, China. Resour Conserv Recycl 49(2):129–46.

ISO (International Organization of Standardization). 2006. Environmental management – life cycle assessment – requirements and guidelines. ISO 14044.

ISO. 2000. Environmental management – life cycle assessment – life cycle impact assessment. ISO 14042.

Kirkeby JT, Birgisdottir H, Hansen TL, Christensen TH, Bhander GS, Mauschild M. 2006. Evaluation of environmental impacts from municipal solid waste management in the municipality of Aarhus, Denmark (EASEWASTE. Waste Manage Res 24:16–26.

Maclaren VW. 2004. Waste management: Integrated approaches. In: Bruce Mitchell, editor. Resource and environmental management in Canada: addressing conflict and uncertainty. Toronto, Canada: Oxford University Press.

McDougall F, White P, Franke M, Hindle P. 2001. Integrated solid waste management: a life cycle inventory. Second Edition. Oxford, United Kingdom: Blackwell Publishing.

Mendes MR, Aramaki T, Hanaki K. 2004. Comparison of the environmental impact of incineration and landfilling in Sao Paulo City as determined by LCA. Resour Conserv Recycl 41:47–63.

Moberg Å, Finnveden G, Johansson J, Lind P. 2005. Life cycle assessment of energy from solid waste - part 2: landfilling compared to other treatment methods. J Clean Prod 13(3):231–40.

Morris J. 2005. Comparative LCAs for curbside recycling versus either landfilling or incineration with energy recovery. Int J LCA 10(4):273–84.

Özeler D, Yetis U, Deminer GN. 2006. Life cycle assessment of municipal solid waste management methods: Ankara case study. Environ Int 32(3):405–11.

Reap J, Roman F, Duncan S, Bras B. 2008. A survey of unresolved problems in life cycle assessment. Part 2: impact assessment and interpretation. Int J LCA 13:374–88.

Rodriguez-Iglesias J, Maranon E, Catrillon L, Riestra P, Sastre H. 2003. Life cycle analysis of municipal solid waste management possibilities in Asturias, Spain. Waste Manage Res 21:535–48.

Shmelev SE, Powell JR. 2006. Ecological–economic modelling for strategic regional waste management systems. Ecolog Econ 59(1):115–30.

Solano E, Ranjithan SR, Barlaz MA, Brill ED. 2002a. Life-cycle-based solid waste management. I: model development. J Environ Eng 128(10):981–92.

Solano E, Dumas RD, Harrison KW, Ranjithan SR, Barlaz MA, Brill ED. 2002b. Life- cycle based solid waste management. II: illustrative applications. J Environ Eng 128(10):993-1005.

Tan RBH, Khoo HH. 2006. Impact assessment of waste management options in Singapore. J Air Waste Manage Assoc 56(3):244–54.

Waste & Resources Action Programme (WRAP). 2006. Environmental benefits of recycling. An international review of life cycle comparisons for key materials in the UK recycling sector. Available at http://www.wrap.org.uk/downloads/ Recycling_LCA_Report_Sept_2006_-_Final.1e36c26c.pdf. Accessed on 2009 01 23.

Winkler J. 2005. Comparative evaluation of life cycle assessment models for solid waste management. Int J LCA 9(6):156–7.

Winkler J, Bilitewski B. 2007. Comparative evaluation of life cycle assessment models for solid waste management. Waste Manage 27:1021–31.

PAPER 2

The incorporation of waste prevention activities into life cycle assessments of municipal solid waste management systems:

Methodological issues

3.1 Introduction

Life cycle assessments (LCAs) of municipal solid waste (MSW) management systems are undertaken to “optimise the infrastructure system for managing a given amount and composition of waste” (Coleman et al. 2003) and are often used to evaluate the validity of the waste hierarchy in identifying those waste management techniques with the lowest environmental emissions and impacts. Waste prevention and product reuse, the first two components of the waste hierarchy, respectively, are usually omitted from these evaluations (Ekvall et al. 2007). Municipal solid waste managers who focus entirely on the ‘end of life’ stage of waste management, that is, once the waste has already been generated, will not succeed in curtailing per capita waste generation, which has increased by 35% in OECD countries since 1980 (OECD 2007). Municipalities have had little control over the policies and interventions that promote waste prevention and product reuse (McKerlie et al. 2006) because these policies are usually the purview of higher levels of government. Recently, cities such as Toronto, Canada have nonetheless attempted to enact policies such as financial incentives for citizens to bring reusable mugs and shopping bags to stores (City of Toronto 2008), in order to reduce waste generation and promote product reuse. Ideally, producers of material goods attempt to design and/or select product systems that minimize environmental burdens, including solid waste, over the product life cycle. However, MSW managers have a different perspective, focusing on the burden of particular products or waste streams on the MSW management system as a whole, a view that is not captured in traditional product LCAs. MSW managers commonly deal with the system-wide effects of changes in processing efficiency and cost, the quantity of material residue generated during sorting and treatment, and the contamination levels of material feedstock for recycling (e.g., Lantz 2008). Although certain technologies, such as those used to improve the sorting of recyclable materials, can be applied to address these MSW management issues, the targeted prevention of particular MSW streams can also be considered. Integrated waste management, commonly perceived as the current standard of practice for waste management, is deemed to address waste treatment, prevention and 44

reuse (McDougall et al. 2001). However, the LCA of MSW management and the theme of waste prevention have tended to remain isolated from one another in the published literature. Exceptions include Coleman et al. (2003), Ekvall et al. (2007) and Olofsson et al. (2004), which allude to the potential for LCAs of MSW to be capable of evaluating waste prevention activities (WPAs) along with waste treatments. The apparent aversion to considering waste prevention as a form of waste management in LCAs of MSW could perhaps be attributable to the difficulty in answering the question: “Can one manage waste that has not been generated?” The fact that this issue has been alluded to in academic publications suggests that the answer to this question is in the affirmative. Most importantly, the WPA is a management process that is not applied to absent waste. The absence of waste is a consequence of the WPA. Therefore, the LCA practitioner can make the apparent philosophical leap to regard waste prevention as functionally equivalent to waste treatments in multi-material MSW LCAs. Waste prevention has long been a subject of academic interest. Several authors have examined policies to encourage waste prevention behaviour by households and industries (e.g., De Young et al. 1993). Of particular importance is the paper by Salhofer et al. (2008), which produces estimates of the waste prevention potential of five different measures (addressing advertising material, beverage packaging, diapers, food waste, and ‘big events’), using Vienna, Austria as a case study. Examples of the waste prevention measures considered include the refusal of unsolicited advertising and the replacement of one-way packaging with refillable packaging. The authors estimate that each measure is capable of producing an approximate 10% reduction in the size of the relevant waste stream. With the debatable exception of waste prevention through the refusal of advertising material, the estimates by Salhofer et al. (2008) are based upon the proviso that a reduction in the unit mass of a product does not decrease the consumption of the service(s) provided by that product. ‘Dematerialization,’ is an expression that is commonly used to represent this form of waste prevention, and has been defined by Van Der Voet et al. (2004) as the “process of fulfilling society’s functions with a decreasing use of material resources over time.” The residents of a municipality in which a WPA through dematerialization is undertaken could experience no apparent reduction in standard of living. 45

It seems likely that published LCAs of MSW rarely compare scenarios that manage different quantities of MSW because the scenarios would be subject to different functional units, thus violating the ISO 14044 international standard which states the requirements and guidelines for LCA (ISO 2006). One possible exception would occur if the functional unit no longer depicts a fixed quantity of MSW managed (e.g., the tonnage of MSW treated per year), but simply the amount of MSW generated in a particular municipality or region, which is assumed to vary in each scenario. In Cleary’s (2009 / Chapter 2) review of 20 published LCAs of multi-material MSW management systems, only one LCA (Rodriguez-Iglesias et al. 2003) compares MSW management scenarios that have different total quantities of MSW collected from the same population, although LCAs by Mohareb et al. (2008) and Olofsson et al. (2004) also include scenarios with different MSW quantities. Rodriguez-Iglesias et al. (2003) and Mohareb et al. (2008) neither account for the upstream benefits of waste prevention, nor address whether or not the types of product services supplied to the population differ for each MSW management scenario. In contrast, Olofsson et al. (2004) do account for the upstream benefits, and employ a consequential LCA methodology (see Section 3.4.2) inasmuch as the effects of waste prevention on markets for recyclable materials are addressed. Similar to Rodriguez-Iglesias et al. (2003) and Mohareb et al. (2008), the scenario comparisons in Olofsson et al. (2003) are not functionally equivalent in terms of product services for the population. None of these studies attribute any potential environmental burdens to the implementation of a WPA. Waste managers and policy developers have lacked an effective tool to help with the evaluation of waste prevention activities in LCAs of MSW, with the significant exception of the U.S. Environmental Protection Agency’s (EPA’s) WAste Reduction Model (WARM). WARM can provide the user with life cycle data on a waste management system’s energy and GHG savings which result from the source reduction of 34 different materials or categories of materials (US EPA 2006). WARM estimates the effects of waste prevention by decreasing the waste inputs into the waste management system and subtracting from the reference scenario the emissions from the product life cycle of the prevented MSW. The LCA practitioner defines the level of waste prevention desired for each scenario, without needing to maintain an equivalent set of product 46 services for the population. For example, the LCA practitioner can compare the life cycle emissions of the reference MSW management scenario with those of a scenario in which the MSW from the disposal of 1000 computers is eliminated. Also of potential significance is WARM’s omission of the effects of each WPA on the treatment of the residual MSW streams.

3.2 Research objectives

In light of the claim by Ekvall et al. (2007) that the traditional LCA model used for MSW management systems is inadequate to address changes in the quantity of waste resulting from WPAs, this paper addresses (1) how the LCA practitioner can incorporate the effects of WPAs in an LCA of a waste management system; and (2) how, in such an LCA of waste, the LCA practitioner can compare waste prevention and treatment activities on a functionally equivalent basis. Section 3.3 of this article briefly reviews the various forms of waste prevention. Section 3.4 examines the existing capabilities of product and waste LCAs, as well as consequential LCA, to address waste prevention, relative to their respective system boundaries. Section 3.5 introduces the Waste Management And Prevention (WasteMAP) life cycle assessment model, which I conceived as a model for undertaking an attributional LCA of MSW with the capability of evaluating the environmental performance of MSW management scenarios incorporating all of the components of the hierarchy of waste management, including waste prevention, product reuse, recycling, and the various forms of waste treatment.

3.3 Types of waste prevention activities

Waste prevention refers to activities undertaken to reduce the mass, volume or toxicity of products or materials consumed, and later discarded, through changes to their “design, manufacture, purchase, or use” (US EPA 1999). The hierarchy of waste management identifies waste minimization or prevention as the most desirable form of waste management, relative to its environmental performance, followed by product reuse, recycling, energy recovery and landfilling (Price and Joseph 2000). Although listed 47

separately in the hierarchy of waste management, product reuse can be considered a form of waste prevention since it almost always results in a reduction in the amount of waste requiring collection (Laner and Rechberger 2009). References to waste prevention in this paper also encompass product reuse activities. WPAs may be undertaken by consumers, producers, and/or MSW managers and can be regulated or otherwise facilitated by government and industry through pricing adjustments, regulation and other means (Salholfer et al. 2008). Consumers can reduce their demand for goods through their purchasing decisions and by using goods more productively. Producers can change the design of a product (e.g., the amount and type of materials required, replacement life and service life) (Cooper 2005, McKerlie et al. 2006) and the production system, resulting in lower environmental emissions, while maintaining product performance. Table 3 describes and provides examples of the various types of WPA. It also lists their properties both in terms of the potential effects of each WPA type on product service(s), and on the presence of alternate product system(s) that would generate additional MSW for treatment. WPAs in accordance with the US EPA’s definition of waste prevention, which explicitly assumes that WPAs reduce waste generation, are listed and classified as WPA types 1-6 in Table 3. WPA type 1 characterises the decrease in the amount of waste generated as a reduction in the quantity of material consumed, without product service substitution. WPA types 2-6 represent dematerialization, which ensures that less MSW is generated, without decreasing the functional outputs associated with the materials entering the MSW management system. MSW can also be prevented at the collection stage of the waste management life cycle. On-property residential waste treatment and the storage of waste products and materials (WPA types 7 and 8) may be considered forms of waste prevention, although it is perhaps more justifiable to consider them as forms of waste diversion since the waste is produced, although it will not be collected for the MSW treatment system. Since waste is generated, there is no effect on the quantity of product services supplied (see Table 3). Product storage, while delaying the treatment of waste, increases the future material liability of a MSW treatment system. This liability places importance on the LCA practitioner’s assumptions that pertain to discount rates and future MSW collection and 48 treatment capabilities. Data on the “hibernating” (stored) portion of products and materials are rare (Brunner 2004).

Table 3 The properties of each type of waste prevention activity

Type of waste prevention Effect of WPA type on Presence of alternate product Example(s) activity (WPA) product service(s) system(s) that contribute additional MSW for treatment WPA-1 Reduction in material Reduction in quantity of No -reduced generation of “junk mail” consumption without product product services (no service substitution substitute product services provided) Dematerialization WPA-2 Reuse of a disposable Substitution of No -reuse of a disposable shopping bag good functionally equivalent product services WPA-3 Substitution of a Substitution of Yes (capital good) -drying of hands by means of hand service, provided by a capital functionally equivalent dryers instead of hand towels good, for a disposable good product services -drinking water supplied from water faucets instead of bottles -newspaper articles available online instead of printed on newsprint WPA-4 Substitution of a Substitution of Yes (substitute reusable good) -substitution of refillable glass wine reusable good for a functionally equivalent bottles for disposable ones disposable one product services -substitution of reusable shopping bags for disposable ones WPA-5 Lightweighting of a Substitution of Yes (substitute disposable good) -substitution of lightweight plastic good functionally equivalent containers for glass ones (both product services containers are single use) WPA-6 Lengthening the Substitution of Yes (substitute durable good) -increasing the useful lifespan of a useful lifespan of a durable functionally equivalent refrigerator through improved good product services design Waste prevention at collection WPA-7 On-property No effect No -backyard composting residential waste treatment -grasscycling

WPA-8 Storage of waste No effect No -storage of obsolete appliances products and materials

3.4 Attributional and consequential approaches to waste prevention and LCA

A reduction in the MSW inputs to the waste management system is inherent to the concept of waste prevention. Waste would have been generated in the conventional manner had it not been for the employment of a particular decision to undertake waste prevention. From this perspective, it would appear that the consequential LCA method is the only approach that is appropriate to evaluate the impacts from reducing MSW inputs. However, within the hierarchy of waste management, waste prevention is considered a form of waste management that is functionally equivalent to others such as landfilling 49

and recycling. From this alternative viewpoint, it is possible to employ the attributional method when undertaking an LCA of MSW that incorporates WPAs.

3.4.1 The attributional approach

The attributional LCA is used to describe a system and its environmental exchanges (Rebitzer et al. 2004), within a pre-set system boundary. It is classified into the traditional product and waste management forms, each having different system boundaries. Both forms possess deficiencies in addressing WPAs.

3.4.1.1 The traditional product LCA

Traditional product LCAs can play an important role in evaluating the net environmental performance of WPAs and are pursued to “optimise a specific product life cycle” (Coleman et al. 2003). They can be used to estimate the differences in the environmental emissions or impacts between product systems, one generating less waste than the other, per unit of functional output supplied. A product system, which is a “collection of unit processes with elementary and product flows, performing one or more defined functions, and which models the life cycle of a product” (ISO 2006), commonly includes raw material extraction, product manufacture, distribution, use and disposal (Coleman et al. 2003). A product LCA addresses the waste management stage only for the product studied. It does not traditionally account for the potential effects of the removal or reduction in size of a waste stream on the entire MSW management system. Thus, it would omit potentially significant effects from the prevention of numerous “problematic materials” that are known to create collection, sorting and processing difficulties that detract from the efficiency of MSW management, and reduce the quality of recycled material. Examples of problematic materials include polylactide (PLA) biodegradable plastics, opaque polyethylene terephthalate (PET) containers and waxed cardboard (Solid Waste Management & Recycling 2008). 50

Recent examples of product LCAs that compare product systems which generate different quantities of waste include Aumônier and Collins (2005), who evaluate the environmental performance of reusable and disposable diapers, and Humbert et al. (2009), who compare plastic and glass packaging for baby food. It is uncommon for product LCAs to provide estimates of the net environmental emissions per tonne of waste prevention (a consequential approach). Instead, the emissions are compared for the different means to supply the same functional unit which reflects the primary use of the good.

3.4.1.2 The traditional LCA of MSW

A traditional LCA of MSW, encompassing waste collection, transportation, sorting and treatment until inert or recycled, can be used to evaluate the life cycle impacts of a waste treatment system in which a particular waste stream is eliminated or reduced in size due to a WPA. It omits the net upstream impacts from implementing the WPA, and the possible substitute product system(s) necessary to maintain an equivalent level of product services to the population under each MSW management scenario. Ideally, all investigated product systems for an LCA begin at the same point – raw material extraction. However, Buttol et al. (2007) claim that “all life cycle stages prior to the product becoming waste can be omitted if they are common to all the subsequent waste management options.” This curtailment of the LCA system boundary, also known as the zero burden approach, simplifies the assessment and allows the LCA to focus on waste treatment. Indeed, Wilson (2002) regards the LCA of MSW as an “end of pipe” model because it cannot address waste minimization or levels of material consumption. A traditional LCA of MSW can incorporate waste prevention by simply excluding the MSW that is prevented, provided that the following criteria are met: (1) the LCA results for the MSW management system are not compared with those of an alternative waste management scenario which addresses a different quantity of waste; and (2) there are no significant additional and/or avoided upstream environmental burdens caused by the implementation of the WPA (e.g., grasscycling). 51

If the first criterion is violated, the MSW management scenarios will not use identical functional units. If the second criterion is not met, the system boundary of the traditional LCA of MSW will exclude the significant environmental emissions from the unit processes (see Appendix 1 for definition) comprising the product system(s) targeted for prevention.

3.4.2 The consequential LCA

Should the objective of the LCA be the accounting of the economy-wide effects of implementing a WPA, consequential LCA is the most appropriate method of analysis. Although the 2006 International Organization for Standardization (ISO) requirements and guidelines for LCA do not recognize the methodology associated with consequential LCA, this LCA type has been described in numerous publications, such as Ekvall and Weidema (2004), and Rebitzer et al. (2004). A consequential LCA is “a model of causal relationships originating at the decision at hand” (Ekvall and Weidema 2004), addressing the economy-wide effects of a change in the functional outputs and inputs on material and energy flows to and from the environment (Curran et al. 2005). Thus, it has a much larger system boundary than an attributional LCA because it also addresses significant flows outside of the MSW management life cycle. Unlike the attributional LCA, this method addresses the marginal effect of a change. There is no need to include within the system boundary those unit processes that would not be affected by the WPA. The functional unit of a consequential LCA would be the amount of waste prevention one intends to undertake.

3.5 The WasteMAP life cycle assessment

In this paper, a conceptual model called the Waste Management and Prevention (WasteMAP) is proposed to facilitate the comparison of MSW management scenarios incorporating waste prevention and the various methods of waste treatment. The amount of MSW managed through treatment and/or a WPA, as well as the functional output of the products that would become MSW, defined by primary and secondary functional 52 units (see Sections 3.5.2.1 and 3.5.2.2); remain constant in all comparative waste management scenarios. The WasteMAP LCA model is attributional inasmuch as its objective is the comparison of the environmental performance of various MSW management scenarios. Unlike WasteMAP, none of the LCA methods described in Section 3.4 can be used to evaluate multi-material MSW management scenarios incorporating all of the components of the hierarchy of waste management. In the WasteMAP LCA, waste prevention through dematerialization (WPA types 2-6 from Table 3) is considered a functional equivalent of MSW treatment and disposal methods, such as incineration and landfilling. It would not be appropriate to consider waste prevention through reduced consumption (WPA-1) as a functionally equivalent method of managing MSW because waste treatments do not affect the composition and magnitude of product services supplied to the population by waste-generating product systems. Moreover, WPAs of types 7 and 8 defined in Table 3 may require the incorporation of additional product systems within the WasteMAP system boundary (e.g., a backyard composter). The following sections describe the methodological characteristics of the WasteMAP LCA, including the system boundary, the functional units, the functional equivalence of product systems, waste flow and environmental emission accounting.

3.5.1 System boundary

System expansion permits the WasteMAP LCA to attain characteristics of both the product LCA system boundary and that of the LCA of MSW. Since there are numerous examples of WPAs which are not influenced by the actions of MSW managers, it is neither feasible nor desirable to account for all of them within the system boundary of the WasteMAP LCA of MSW. Thus, it is necessary to distinguish between WPAs that are implicitly included in MSW management scenarios, and “additional WPAs,” which are explicitly included within the WasteMAP system boundary, with their net upstream and downstream benefits taken into account. A simplified depiction of the system boundary of the WasteMAP LCA is illustrated in Figure 8, which uses a material 53

substitution WPA as an example. This system boundary varies with the types of WPAs included in each scenario (Table 3). WasteMAP’s upstream component has a system boundary similar to a product LCA (raw material extraction, processing, manufacturing, transportation, and use), although excluding the waste treatment stages. Its downstream component includes the collection, transportation, sorting, treatment and disposal of the many MSW streams that enter the MSW management system. The upstream component of WasteMAP illustrated in Figure 8 addresses only the product systems, excluding the waste treatment stages, which are potentially affected by WPAs: the targeted product systems (TPSs) and the alternate product systems (APSs). When applying the “avoided burden” approach (Frischknecht and Jungbluth 2007), the former refers to the product systems that are subtracted from the total MSW subject to treatment, while the latter depicts the product systems that may need to be added to the total MSW. Accounting for the impacts of TPSs and APSs can be a very complex endeavour, as they may include product systems for new capital, products, and services. MSW and upstream wastes can potentially be generated as a direct consequence of implementing a WPA which fully maintains the types of product services supplied to the population. The differing widths of the TPS and APS in Figure 8 are intended to illustrate that the amount of waste removed from the MSW treatment system must be greater than the amount added to it. If an LCA of MSW would evaluate waste management scenarios with different quantities of waste, such as those which account for WPAs, Ekvall et al. (2007) claim: “It is reasonable to demand that such studies include the environmental burdens associated with the production of all the materials that eventually become waste.” Although acquiescing to this demand would produce a greatly expanded system boundary, the LCA results would be far more difficult for the waste manager or policy- maker to interpret. Waste prevention and waste management processes would not be isolated from the numerous product systems unaffected by the waste prevention activities. The LCA practitioner can minimize data requirements by maintaining the zero burden assumption for those product systems unaffected by WPAs and applying system expansion to account for the avoided and additional burdens of the TPSs and APSs. The primary methodological complication of this approach is evident in cases in which the 54

TPSs or APSs possess significant co-products, possibly resulting in an asymmetry. In order to ensure that the single product output of interest, measured by the secondary functional unit (see Section 3.5.2.2), remains isolated, the LCA practitioner can subtract compensatory single output processes. An asymmetry ensues if the sum of the impacts from the individual processes associated with the product output of interest and the co- products is not identical to the impacts from the original multi-output processes of the TPSs and APSs. One method to avoid such an asymmetry is to apply an allocation.

Figure 8 Basic system boundary of the WasteMAP LCA. Note: Some graphical elements of Figure 8 were derived from McDougall et al. (2001)

3.5.2 Functional units

The functional unit provides a common basis for LCA result interpretations and comparisons (Rebitzer et al. 2004). Although this concept is defined as the “quantified performance of a product system for use as a reference unit” (ISO 2006), the ISO’s 55

definition requires the substitution of ‘waste management’ for ‘product’ in an LCA of MSW. The functional unit of a product LCA is defined by the output of the system, whereas that of an LCA of MSW is defined in terms of the system’s input (McDougall et al. 2001). Functional units for the WasteMAP LCA are analogous to those of product and waste management LCAs. Primary and secondary functional units ensure both a fixed amount of MSW managed in each scenario, as well as identical reference flows of functionally equivalent product services.

3.5.2.1 Primary functional unit

The primary functional unit of the WasteMAP LCA is the amount (mass or volume) of material addressed by the MSW management system on an annual basis. It is identical for all MSW management scenarios and is equal to the sum of the upstream primary functional unit (UPFU) and the downstream primary functional unit (DPFU). The UPFU is defined as the net amount of material left out of the MSW treatment system due to WPAs, whereas the DPFU tracks the amount of MSW collected and treated under each scenario. Since waste management scenarios that are considered functionally equivalent can differ in the amount of waste prevention undertaken, they need not have identical UPFUs and DPFUs. The amount of solid waste generated upstream in product life cycles, including the mining overburden removed during raw material extraction, tends to greatly surpass the post-consumer waste (Washington State Department of Ecology 2003). The decision of whether or not to incorporate these wastes in the primary functional unit could be of high significance to the LCA results. Thus, the LCA practitioner could include in the primary functional unit not only the residential waste managed through prevention and treatment, but also the upstream changes in the institutional, commercial and industrial waste resulting from the WPA(s) (see Section 3.5.3). When considering the mass equivalence of each waste management scenario, all of these waste types would be taken into account. Consequently, some WPAs, deemed equivalent in terms of the total mass of waste prevented, may not result in the prevention of the same amount of residential waste. The MSW management scenarios would not be equivalent from the perspective of 56

the residential waste manager because the amount of waste collected and treated for each would differ.

3.5.2.2 Secondary functional unit

Product services supplied to municipal residents can often be delivered through alternative means and thereby reduce total MSW generation. In the WasteMAP LCA, the product systems that provide these services are depicted by the TPS and the APS (see Section 3.5.1). Analogous to functional units in product LCAs, WasteMAP’s secondary functional units are used to ensure that MSW management scenarios subject to comparison will supply functionally equivalent product services to the residents of the municipality. The reference flows of the product systems added and removed from the MSW treatment system must also remain identical within each MSW management scenario. Secondary functional units are only applicable when scenario comparisons include waste prevention in the form of dematerialization (i.e., WPAs 2-6 from Table 3). For example, a secondary functional unit can address the function of supplying packaged juice to the residents of a municipality in a particular year. The means of delivering such a product service (e.g., a plastic or glass juice container), can differ. However, the reference flows of the TPSs and APSs that depict the quantity of packaged juice supplied to the population in each respective MSW management scenario would be equal. It is not necessary for the LCA practitioner to define secondary functional units for the product functions associated with each MSW stream, but only to those wastes which are affected by WPA types 2-6. Secondary functional units cannot be applicable to waste management scenarios that incorporate WPA-1, since there would be no replacement product service provided. For this WPA type, MSW management scenarios cannot be compared with the reference MSW management scenario on a functionally equivalent basis. There are two exceptions in which MSW scenarios under WasteMAP do not require a secondary functional unit to ensure the functional equivalence of product services: (1) the TPSs supply product services that are deemed unwanted by certain 57 segments of the population (e.g., unaddressed advertising); and (2) the removal of waste from the MSW treatment system does not decrease the quantity of product services supplied to consumers (e.g., grasscycling). The properties of a product are related to its functionality, technical quality (stability, durability, ease of maintenance), additional services rendered, aesthetics, image, economic costs and specific environmental properties (Weidema et al. 2004). Therefore, consumers may not always receive identical or equivalent product services when implementing waste prevention through product substitution (Weidema et al. 2004). For example, the lightweighting of glass bottles could result in an increase in the frequency of bottle breakage. Similar to the functional unit in most comparative product LCAs, only the product property of significance to the study is measured by the secondary functional unit, since not all of the properties are used or are necessary to carry out the product service required of it (Cooper 2003). In assessing the waste prevention potential of WPAs, the LCA practitioner has to evaluate whether or not the consumer views the replacement product service as a reasonable substitute for the original. He/she must also account for the effect of consumer habits on the performance and lifetime of each product system (Günther and Langowski 1997). For example, if the primary reason for a consumer to replace a particular durable good is aesthetic, an increase in product durability will have little or no impact on the amount of waste generated. Should the targeted product system encompass more than one significant function, the effects of the additional functions could be subjected to an allocation procedure, or a boundary extension (Cooper 2003).

3.5.3 Waste flows

The net flows of waste relative to the default quantity of residential waste generated may be calculated through the application of equations 1-3 below. Default residential waste generation in the reference WasteMAP scenario is the amount of residential waste (mass or volume) generated in the municipality of interest in the absence of additional WPAs, classified by product/material type, with the types of units used to depict mass or volume left to the discretion of the LCA practitioner. In contrast, 58 net residential waste generation (RNET) takes into account the effects of additional WPAs on the default residential waste generated, usually on an annual basis. For each scenario, the cumulative net residential waste generation, irrespective of composition, is calculated in Equation 1, assuming a total of “n” WPAs:

n n RNET =RAPSWPA – RTPSWPA (1) WPA=1 WPA=1 where RAPS is the residential waste generation potentially added to the MSW treatment system due to WPAs, while RTPS is the residential waste subtracted from the MSW treatment system. In order for net residential waste prevention to take place, RNET must be less than zero. Although the WasteMAP LCA is conceived as a model for residential MSW management systems, the LCA practitioner may wish to account for the potential of WPAs to affect the production of upstream industrial, commercial and institutional (ICI) solid waste, including the eventual disposal of capital equipment. The cumulative net ICI waste generation (ICINET), irrespective of composition, is calculated by Equation 2, assuming a total of “n” WPAs:

n n ICINET = ICIAPSWPA – ICITPSWPA (2) WPA=1 WPA=1 where ICIAPS is the additional ICI waste that is potentially generated upstream, and ICITPS is the upstream ICI waste that would have resulted from the production of the product system(s) subtracted from the MSW stream due to the WPAs. Net waste generation (WNET), which takes into account the flows of residential, industrial, commercial and institutional waste, is calculated as follows, assuming that the WPAs in equations 1 and 2 are identical:

WNET = RNET + ICINET (3)

3.5.4 Environmental emissions

Equation 4 is used to calculate the environmental emissions of a WasteMAP LCA scenario, assuming that the implementation of “n” WPAs has no effects on the waste treatment system for the remaining MSW:

n n WMP = REF – PSWPA + APSWPA (4) WPA=1 WPA=1

where WMP represents a particular environmental emission from a WasteMAP LCA scenario that includes WPAs, REF symbolizes the emission from the reference waste management scenario that lacks WPAs, TPS signifies the environmental emission from each reference flow of product systems subtracted from the MSW treatment system, and APS represents the environmental emission from each reference flow of alternate product systems that generate less waste. Equation 4 does not always represent the effect of implementing WPAs because WPAs change the composition and quantity of waste collected for treatment, which can generate effects throughout the MSW management system. Equation 4 is representative if the MSW streams associated with the TPSs and APSs are collected, sorted and treated separately from the remaining MSW streams, such as used containers collected through a deposit-return system. The use of Equation 4 is inappropriate if the presence of WPAs results in significant effects upon: (1) the efficiency of the collection, sorting, processing, treatment and disposal components of the MSW management system; (2) the overall contamination levels of collected materials for source-separated organics and recycling programs; (3) the physical and/or chemical reactions between the remaining MSW streams undergoing treatment; and (4) the quality of the processed waste material. Many of these effects are associated with the presence of so-called “problematic materials” for recycling programs (Waste Diversion Ontario 2009) and source-separated organics treatment programs. 60

Equation 5 estimates the net change in downstream emissions from the management of the remaining MSW (NETDOWN):

n n NETDOWN = DOWN – (REF – DownTPSWPA + DownAPSWPA) (5) WPA=1 WPA=1

where DOWN, which can be derived through a traditional MSW LCA, depicts the downstream environmental emission from a MSW management scenario that includes WPAs; DownTPS represents the downstream environmental emission from each reference flow of product systems targeted for removal; and DownAPS signifies the downstream environmental emission from each reference flow of alternate product systems that generate less waste. If NETDOWN is insignificant, the LCA practitioner can use Equation 4 to account for the impacts of waste management scenarios. The use of Equation 6, defined below, is appropriate if the result from Equation 5 is significant, indicating that there are substantial downstream impacts of the WPAs on the management of the MSW remaining in the waste treatment system, such as the impacts from preventing the generation of “problematic materials.” Equation 6 is defined as follows:

n n WMP = DOWN – UpTPSWPA + UpAPSWPA (6) WPA=1 WPA=1

where, as in Equation 4, WMP represents a particular environmental emission from a WasteMAP LCA scenario that includes WPAs, UpTPS signifies the upstream environmental emission from the product systems targeted for removal, and UpAPS represents the upstream environmental emission from the alternate product systems that generate less waste. For managing the same MSW, Equation 7 of the WasteMAP LCA is used to estimate the net environmental emissions attributed to the implementation of waste prevention activities (WP):

n n WP = DOWN – REF – UpTPSWPA + UpAPSWPA (7) WPA=1 WPA=1

Equation 7 not only allocates the net change in upstream emissions to the WPAs, but also the net downstream environmental emissions, which would include possible changes in the management of the MSW not subject to waste prevention.

3.6 Discussion

When selecting a method to evaluate the environmental performance of residential waste management systems which include waste treatment, diversion and prevention activities, the LCA practitioner has several options. Consequential LCA is appropriate when focusing only on the life cycle effects of a change in the quantity of waste managed. Product LCAs are justified if the focus of the LCA is to compare the emissions from two product systems providing equivalent product services, with one generating less residential waste than the other. Alternatively, the LCA practitioner can apply the WasteMAP LCA model to compare within one system boundary all of the components of the waste management hierarchy, including WPAs and waste treatments, so that they are regarded as functional equivalents in managing MSW. This would permit the LCA practitioner to identify in the LCA results the burdens and avoided burdens attributed to waste prevention, recycling, biological and thermal treatments, as well as landfilling, within a particular MSW management system. Should the product systems targeted for prevention by the MSW manager supply different product services measured with one or more secondary functional units, WasteMAP’s Equation 4 is appropriate. However, if the WPAs have significant impacts on the management of the remaining waste streams, WasteMAP’s Equation 6 should be employed. The WasteMAP LCA uses system expansion similar to Olofsson et al. (2004) and the US EPA’s Waste Reduction Model (WARM). However, WasteMAP requires functional equivalency in all compared scenarios and introduces an additional type of functional unit. Its equations address the net emissions attributed to the introduction of 62

WPAs into a waste management system as well as the issue of waste prevention effects on the treatment of the residual MSW streams. The utilization of the WasteMAP model may possess a number of limitations, including the potential to increase the amount of error and uncertainty in the results due to the expansion of system boundaries. Error can also arise from assumptions associated with the functional equivalence of product services, such as assigning overly narrow functionally equivalent comparisons (Reap et al. 2008) for the secondary functional units of the WasteMAP LCA. It may, therefore, be of use to survey the target population to generate an estimate of the percentage of consumers that considers two product systems to be substitutable or a certain product service to be unnecessary. The functional equivalence assumption for a particular WPA may not be appropriate beyond a certain threshold. The LCA practitioner would need to distinguish between the substitutability of product systems and the preference for one product system over another due to differences in the perception of quality, aesthetics or other properties. Some MSW streams are known to be responsible for the contamination of various waste material feedstocks for recycling or for biological treatment. For example, although the biodegradable plastic, PLA, can be recycled, it cannot be combined with other recyclable plastics without harming the existing plastics recycling systems (Franklin Associates 2006). MSW managers may wish to evaluate MSW management scenarios that target the prevention of such waste streams in order to obtain downstream efficiency benefits. The WasteMAP model is also suited for such an evaluation: comparing the environmental performance of the reference MSW management scenario with one that lacks particular waste streams. In advance of conducting an LCA based on WasteMAP, it would be wise to evaluate the waste prevention potential of a MSW management system and the potential effectiveness of various measures to implement WPAs. Additional research could also be undertaken to explore the extent to which the implementation of WPAs has decreasing returns to scale and whether or not there would be a greater waste prevention potential from cities with a larger per capita production of waste.

3.7 Conclusion

While allowing the LCA practitioner to reap the data collection savings associated with streamlining the LCA, fixing the “cradle” of LCAs of MSW at waste generation limits the utility of the LCA to waste managers and policy-makers due to the exclusion of waste prevention activities. System boundary expansion and the introduction of an additional type of functional unit permit the WasteMAP LCA to be applied to a wider array of possible waste management scenarios, including those which incorporate waste prevention. WasteMAP, unlike the traditional LCA of MSW, possesses the capability of evaluating the validity of the hierarchy of waste management, a resilient concept that continues to evince a significant influence on waste management policies.

3.8 References

Aumônier S, Collins M. 2005. Life Cycle Assessment of Disposable and Reusable Nappies in the UK. Environment Agency, Bristol, UK. Available at http://publications.environment-agency.gov.uk/epages/eapublications.storefront/ 4aaa7c480068bf9a273fc0a8029606d5/Product/View/SCHO0505BJCW&2DE&2DE#. Accessed on 2010 07 01.

Brunner P. 2004. Materials Flow Analysis and the Ultimate Sink. J Ind Ecol 8(3):4-7.

Buttol P, Masoni P, Bonoli A, Goldoni S, Belladonna V, Cavazzuti C. 2007. LCA of integrated MSW management systems: case study of the Bologna District. Waste Manage 27(8):1059-1070.

City of Toronto. 2008. Proposed Measures to Reduce In-Store Packaging Waste and Litter, Municipal Hazardous and Special Waste and Plastic Water Bottles. Staff report. City of Toronto Solid Waste Management Services. Ref. No. p:/2008/swms/Nov./019PW, Available at http://www.toronto.ca/legdocs/mmis/ 2008/pw/bgrd/backgroundfile-17097.pdf. Accessed on 2008 07 12.

Cleary J. 2009. Life cycle assessments of municipal solid waste management systems: A comparative analysis of selected peer-reviewed literature. Environ Int 35(8):1256-1266.

Coleman T, Masoni P, Dryer A, McDougall F. 2003. International expert group on life cycle assessment for integrated waste management. Int J LCA 8(3):175-178.

Cooper J. 2003. Specifying functional units and reference flows for comparable alternatives. Int J LCA 8(6):337-349. 64

Cooper T. 2005. Slower Consumption: Reflections on Product Life Spans and the "Throwaway Society." J Ind Ecol 9(1-2):51-68.

Curran M, Mann M, Norris G. 2005. The international workshop on electricity data for life cycle inventories. J Clean Prod 13:853-862.

De Young R, Duncan A, Frank J, Gill N, Rothman S, Shenot J, Shotkin A, Zweizig M. 1993. Promoting Source Reduction Behavior: The Role of Motivational Information. Environ Behav 25:70-85.

Ekvall T, Assefa G, Bjorklund A, Eriksson O, Finnveden G. 2007. What life-cycle assessment does and does not do in assessments of waste management. Waste Manage 27(8):989-996.

Ekvall T, Weidema B. 2004. System Boundaries and Input Data in Consequential Life Cycle Inventory Analysis. Int J LCA 9(3):161-171.

Franklin Associates. 2006. Life cycle inventory of five products produced from polylactide (PLA) and petroleum-based resins: technical report. Prepared for Athena Institute International. Available at http://www.athenasmi.ca/ projects/docs/Plastic_Products_LCA_Summary_Rpt.pdf. Accessed 2010 07 01.

Frischknecht R, Jungbluth N. 2007. Overview and methodology: EcoInvent report no. 1. Swiss Centre for Life Cycle Inventories.

Günther A, Langowski H-C. 1997. Life Cycle Assessment Study on Resilient Floor Coverings. Int J LCA 2(2):73-80.

Humbert S, Rossi V, Margni M, Jolliet O, Loerincik Y. 2009. Life cycle assessment of two baby food packaging alternatives: glass jars vs. plastic pots. Int J LCA 14(2):95-106.

ISO. 2006. Environmental Management – Life Cycle Assessment. Principles and framework. ISO 14040: 2006.

Laner D, Rechberger H. 2009. Quantitative evaluation of waste prevention on the level of small and medium sized enterprises (SMEs). Waste Manage 29:606-613.

Lantz, D. 2008. Mixed Results. Resour Recycl 11-15.

McDougall F, White P, Franke M, Hindle P. 2001. Integrated Solid Waste Management: a Life Cycle Inventory. Second Edition. Oxford, United Kingdom: Blackwell Publishing.

McKerlie K, Knight K, Thorpe B. 2006. Advancing extended producer responsibility in Canada. J Clean Prod 14:616-628. 65

Mohareb A, Warith M, Diaz R. 2008. Modelling greenhouse gas emissions for municipal solid waste management strategies in Ottawa, Ontario, Canada. Resour Conserv Recycl 52:1241–1251.

OECD. 2007. OECD Environmental Data: Compendium 2006/2007. Waste. Paris, France: Environmental Performance and Information Division, OECD Environment Directorate. Working Group on Environmental Information and Outlooks. Available at http://www.oecd.org/dataoecd/60/59/38106368.pdf. Accessed on 2009 03 16.

Olofsson M, Ekvall T, Sundberg J. 2004. Impacts of Swedish waste prevention and the scrap market equilibrium on greenhouse gas emissions. In: M. Olofsson. Improving Model-Based Systems Analysis of Waste Management. PhD thesis, Department of Energy Technology, Chalmers University of Technology, Gothenburg, Sweden.

Price J, Joseph J. 2000. Demand management – a basis for waste policy: a critical review of the applicability of the waste hierarchy in terms of achieving sustainable waste management. Sust Dev 8(2):96-105.

Reap J, Roman F, Duncan S, Bras B. 2008. A survey of unresolved problems in life cycle assessment. Part 1: goal and scope and inventory analysis. Int J LCA 13:290-300.

Rebitzer G, Ekvall T, Frischknecht R, Hunkeler D, Norris G, Rydberg T, Schmidt W, Suh S, Weidema B, Pennington D. 2004. Life cycle assessment Part 1: Framework, goal and scope definition, inventory analysis, and applications. Environ Int 30(5):701-720.

Rodriguez-Iglesias J, Maranon E, Catrillon L, Riestra P, Sastre H. 2003. Life cycle analysis of municipal solid waste management possibilities in Asturias, Spain. Waste Manage Res 21:535-548.

Salhofer S, Obersteiner G, Schneider F, Lebersorger S. 2008. Potentials for the prevention of municipal solid waste. Waste Manage 28:245-259.

Solid Waste Management & Recycling. 2008. The Dirty Dozen: Twelve materials that create problems for recycling plants. Solid Waste Manage & Recycl 13(6):51-52.

US EPA. 2006. Solid Waste Management and Greenhouse Gases: A Life-Cycle Assessment of Emissions and Sinks. 3rd Edition, Available at http://www.epa.gov/climatechange//wycd/waste/downloads/fullreport.pdf. Accessed on 2008 12 26.

US EPA. 1999. National Source Reduction Characterization Report for Municipal Solid Waste in the United States. Washington, D.C.: United States Environmental Protection Agency. 77 pp.

Van Der Voet E, Van Oers L, Nikolic I. 2004. Dematerialization: Not just a matter of weight. J Ind Ecol 8(4):121-137. 66

Washington State Department of Ecology. 2003. Beyond Waste: Waste and Material Flows in Washington. Prepared by Cascadia Consulting and Ross and Associates. Publication number 03-04-028. Available at http://www.ecy.wa.gov/pubs/0304028.pdf. Accessed on 2009 10 10.

Waste Diversion Ontario. 2009. Recommended Process to Identify and Address Printed Papers and Packaging that are Problematic for Recycling Programs. Draft for Consultation. Available at http://www.wdo.ca/files/domain4116/Draft%20Problematic %20Materials %20Process%20Nov%2017%20for%20posting.pdf. Accessed 2009 12 04.

Weidema B, Wenzel H, Petersen C, Hansen K. 2004. The Product, Functional Unit and Reference Flows in LCA. Danish Environmental Protection Agency, Environmental News No. 70.

Wilson E. 2002. Life cycle inventory for Municipal Solid Waste management. Part 2: MSW management scenarios and modeling. Waste Manage Res 20:23-36.

PAPER 3

Life cycle assessments of wine and spirit packaging at the product and the municipal scale:

A Toronto, Canada case study

4.1 Introduction

Packaging alternatives to conventional single use glass bottles for wines and spirits have been introduced to markets around the world and promoted as a means of reducing waste and environmental impacts. The presence of these alternatives is a relatively recent phenomenon, exemplified by the introduction of the first wine sold in an aseptic carton in Ontario, Canada in late 2005 (LCBO 2006). The academic and business communities have demonstrated an interest in the environmental impacts of wine and spirit packaging (e.g., Bengoa et al. 2009; Franklin Associates 2006), and alternatives to conventional glass containers are often marketed as “eco-friendly,” particularly in relation to their contributions to waste prevention (LCBO n.d.). Alternative packaging options for wines and spirits include lightweight glass and polyethylene terephthalate (PET) plastic bottles, and aseptic cartons, among other types of containers (Bengoa et al. 2009). The Liquor Control Board of Ontario (LCBO) has promoted the adoption of lightweight forms of packaging for wines and spirits as a means of implementing its plan to eliminate 10 million kg of solid waste on an annual basis (LCBO 2006). Life cycle assessment (LCA) can be used to account for the net environmental emissions and impacts that result from such initiatives to manage residential waste. One alternative packaging option that the LCBO has not pursued is glass bottle reuse. The bottle refilling alternative, although relatively common for wines in Europe, is rarely undertaken in North America (Leighton 2010). Historically, there has been resistance to the adoption of refillable glass containers for many reasons, such as the initial cost of setting up the system, as well as issues associated with the potential number of refills for each bottle, the efficiency of collection and handling, and the transport distance to the refilling facility (Moody 1977). However, instituting a refillable bottling system for wines and spirits may be feasible for the province of Ontario since it possesses a deposit-return system for such containers and a large indigenous wine industry situated relatively close to the largest market in Ontario (Greater Toronto). The province imports large quantities of bulk wines and spirits that are packaged domestically. Moreover, 69

Ontarians are already familiar with the deposit-return system for refillable beer containers and have an impressive return rate for these containers, reaching 99% (TBS 2009). A number of industry-funded LCAs of alternative forms of wine packaging and bottle closures have been published (e.g., Franklin Associates 2006, PwC/Ecobilan 2008), but I am unaware of any LCAs addressing containers for spirits. Some industry- funded packaging comparisons using LCA have been relatively controversial, at times generating seemingly contradictory results without adequate explanations for the differences. For example, results from an LCA commissioned by a glass manufacturer (Owens-Illinois 2010) indicate that glass containers generate a much smaller carbon footprint than plastic ones, while another LCA commissioned by the plastics industry (PET Resin Association) (Franklin Associates 2009) finds the reverse. When possible, it is important for an LCA to explore the reasons for such substantial differences, such as the role played by the selected system boundaries. Two types of LCA scenarios are compared in this paper. The first set of scenario comparisons is at the scale of an individual 750 ml and one litre package. The second set addresses the life cycle impacts from the packaging required for all wines and spirits consumed by residents of the City of Toronto, outside of commercial establishments, in 2008. The 2008 reference scenario is compared with a functionally-equivalent (in terms of the volumes of packaged wines and spirits supplied) alternative packaging scenario. This alternative scenario, if implemented, would decrease residential waste generation by approximately half (46.9% by mass). An LCA at the municipal scale which addresses the current environmental impacts from wine and spirit packaging, as well as the potential for environmental gains through the adoption of lightweight and refillable packaging, has not been undertaken until now. This municipal context permits the accounting of the actual levels of consumption of a wide array of packaging used for a particular product supplied to municipal residents, as well as the specific characteristics of the municipal solid waste management system (e.g., recycling rates, locations of treatment facilities).

4.2 Research objectives and methodology

The research objectives of this paper are: (1) to evaluate, by means of a life cycle assessment, the net environmental impacts from the packaging of wines and spirits supplied to residents of the City of Toronto, Canada in 2008; (2) to compare these results with an alternative scenario in which lightweight and refillable forms of packaging are used; and (3) to compare the life cycle impacts of five types of individual one litre wine packages and four types of 750 ml spirit packages. Unlike many product LCAs, the LCA scenarios addressed in this paper do not only focus on the environmental emissions and impacts from the life cycle of a particular product. Rather, they focus on the actual sales of the packaged product (in multiple types and sizes of packages) in a particular municipality over a specific period of time. This set of municipal scale LCAs uses commercial geography data (i.e., volume sales of wines and spirits) to identify the reference flows of an agglomerated set of functionally equivalent packaging products. In this case, the input data depict the amounts of primary (containers) and secondary (closures, capsules and labels) packaging used for the wines and spirits consumed by the 2.5 million residents of the City of Toronto, Canada in 2008. Other than conventional single use (CSU) glass containers, four alternative means of packaging wine and spirits are addressed in the LCA: [1] lightweight single use (LSU) glass bottles; [2] refillable glass (RFG) bottles; [3] polyethylene terephthalate (PET) bottles; and [4] aseptic cartons (for wine only). All of these alternatives would reduce residential waste generation because the mass of these containers, per unit volume of liquid contained over the lifespan of the package, is less than for CSU glass bottles. LCAs are also undertaken to compare results for each type of packaging system at the scale of an individual package, using the same Toronto, Canada context. Although the 750 ml container has the highest sales for both wines and spirits, the LCAs of individual wine packages compare only one litre packages because aseptic cartons are unavailable in the 750 ml format. Since no spirits were sold in one litre containers in 2008, the 750 ml package is selected for the LCA comparisons at the scale of the individual spirit package. Only four container types are compared for spirits because spirits are not sold in aseptic cartons. 71

4.2.1 Functional units

The first set of LCA scenarios (Section 4.2.2.1), which addresses individual packages, uses one litre of packaged wine as the functional unit for wine containers, and 750 ml of packaged spirits for spirit containers. For the second set (Section 4.2.2.2), the functional unit depicts the volume of packaged wines and spirits, excluding wines in “bag-in-box” containers, that was consumed domestically (i.e., outside of commercial establishments) by the residents of the City of Toronto in the year 2008. This volume was approximately 21.6 million litres of wine and 11.1 million litres of spirits (see Section 4.3 for calculation procedure). Although various properties of aseptic cartons, PET, LSU and RFG bottles differ from CSU glass containers, they are substitutes in that they are all used to package wines and spirits. In this LCA, the relatively common “bag- in-box” packaging for wine is not considered a replacement for these other containers, since it generally holds a much greater volume of liquid (i.e., 3 or 4 litres). With exceptions noted below, it is assumed that aseptic cartons, PET bottles, LSU and refillable glass bottles provide a product service that is equivalent to CSU glass bottles in the packaging of wines and spirits. However, these replacement containers are not perfect substitutes. Spirits tend to have greater alcohol content than wine and are often consumed over longer periods of time, which can affect packaging requirements. Sparkling wines require packages that contain the wine at relatively high pressures. Other wines, largely more expensive vintages, are often cellared in bottle for a considerable length of time before opening. Unlike glass containers, the aseptic carton and PET bottle are inappropriate for longer term storage of wines and spirits. A 2010 study by l'Institut des Sciences de la Vigne et du Vin of Bordeaux, France observed

adverse effects of PET containers on the oxygen penetration, SO2 content, taste and colour of wine, especially white wine, in as little as three months (ISVV 2010). Furthermore, aseptic cartons are not appropriate for liquids under pressure. Although commissioned by an interested party in 2009, a poll of 1300 North American wine- drinkers observed that glass containers are preferred over other packaging alternatives (Owens-Illinois 2009). This appears to be confirmed by the relatively small volume of wines and spirits sold in alternative containers (see Section 4.2.2.2). 72

4.2.2 Packaging scenarios

4.2.2.1 LCA scenarios for wine and spirit packaging at the scale of an individual package

The LCA scenarios for individual wine packages are designated as follows: one litre of wine packaged in a: (1) one litre conventional single use glass bottle; (2) one litre lightweight single use glass bottle; (3) one litre refillable glass bottle; (4) one litre PET bottle; and (5) one litre aseptic carton (Figure 9). The following individual packaging systems for spirits are also evaluated: (1) 750 ml CSU glass bottle; (2) 750 ml LSU glass bottle; (3) 750 ml PET bottle; and (4) 750 ml refillable glass bottle.

Figure 9 The one litre PET bottle, the one litre tetra prisma aseptic carton, and the one litre conventional single use glass bottle (left to right).

4.2.2.2 LCA scenarios for wine/spirit packaging consumption at the municipal scale

The municipal scale LCA scenarios are defined as follows:

2008 reference scenario (Table 4) The composition of wine and spirit packaging is identical to that supplied to the residents of the City of Toronto in 2008.

Alternative packaging scenario (Table 5) A hypothetical scenario in which: (1) Refillable glass (RFG) bottles are substituted for all 750 ml, 1000 ml, and 1500 ml containers for Canadian and bulk imported wines, as well as 750 and 1140 ml containers for Canadian and bulk imported spirits; (2) Aseptic cartons (ACs) are used to package 10% of the 2008 volume sales of imported packaged wines in 1500 ml containers, and 50% of the 2008 volume sales of imported packaged wines in 1000 ml containers; (3) PET bottles replace 10% of the 2008 volume sales of imported packaged spirits in 750 ml and 1140 ml containers, and 50% of the 2008 volume sales of imported packaged spirits in 200 ml, 375 ml, 1750 ml and the remaining container sizes; and (4) The remaining containers are packaged in lightweight single use (LSU) bottles (assumed to have a 20% lower mass than CSU containers, for all container sizes).

The composition of the wine and spirit packaging used in the 2008 reference scenario is based upon the LCBO 2008 volume sales datasets provided to the author for analysis in order to determine the percentage of containers of a particular volume that were aseptic cartons, glass and PET bottles (LCBO 2008) (Table 4). The percentages are only estimated for those container sizes comprising approximately 1% or more of wine or spirit volume sales. These data are assumed representative of the packaging composition of wines/spirits consumed by residents of the City of Toronto in 2008. Based on a broad review of the dataset figures, wine containers within the “other” category are assumed to have a volume of 375 ml, with spirit containers at 750 ml.

Table 4 Estimated consumption of wines and spirits in Toronto by size and type of container, 2008

Size of container Estimated sales % of wine / % of each container size by type of container2 (ml) in Toronto by spirits contained Aseptic carton Glass bottle PET bottle volume (litres)1 in each packaging size2 WINES 750 1.35*107 55.5 0 99.5 0.5 1000 1.40*106 10.5 35.3 61.5 3.2 1500 6.98 *106 31.1 0.1 99.9 0 2000 2.38*105 1.1 0 100 0 3000 2.24*105 1.0 0 100 0 Other 1.97 *105 0.9 N/A N/A N/A TOTAL WINES 2.25*107 100% (excl. BIB) SPIRITS 74

200 1.85*105 1.6 0 84.8 15.2 375 1.05 *106 9.1 0 71.9 28.1 750 4.61*106 39.8 0 97.1 2.9 1140 3.89*106 33.6 0 97.3 2.7 1750 1.74*106 15.0 0 63.1 36.9 Other 1.04 *105 0.9 N/A N/A N/A TOTAL SPIRITS 1.16*107 100% 1 The estimated sales in the City of Toronto (by volume) are based upon Ontario sales data supplied by the Association of Canadian Distillers (2008) and the Canadian Vintners Association (2008). The Ontario figures have been multiplied by the percentage of the provincial population that resided in the City of Toronto in 2006 (19.76%) (City of Toronto 2007, Statistics Canada 2006). 2 The percentages of wines/spirits contained in each packaging size and the percentages of each container size by type of container (excluding bag-in-box containers) are derived from 2008 LCBO volume sales data. Wine and spirit coolers, as well as ciders, are excluded from these figures.

The results displayed in Table 4 show the relative predominance of 750 ml and 1500 ml containers to package wines, as well as 750 ml and 1140 ml bottles to package spirits. These data also reveal the much greater popularity of glass containers over the alternatives for packaging both wines and spirits. Table 5 displays the percentage assumptions in the alternative packaging scenario, which are intended to represent plausible levels of the market penetration of various alternative packaging systems. These alternative packaging substitutions would result in a substantial reduction in the tonnage of packaging waste generated (46.9%).

Table 5 Estimated consumption of non-bag-in-box wines and spirits in Toronto by size and type of container in the alternative packaging scenario

Size of % of each container size by type of container container (ml) Aseptic carton Lightweight glass bottle PET bottle Refillable glass bottle WINES 750 0.00 72.1 0 27.9 1000 36.0 36.0 0 27.9 1500 7.2 64.8 0 27.9 2000 0 100 0 0 3000 0 100 0 0 Other 0 100 0 0 SPIRITS 200 0 50.0 50.0 0 375 0 50.0 50.0 0 750 0 36.1 4.0 59.9 1140 0 36.1 4.0 59.9 1750 0 50.0 50.0 0 Other 0 50.0 50.0 0

A number of articles in waste management publications (e.g., Valiante 2007) have discussed the feasibility of introducing a refillable wine bottle program to Ontario. 75

Moreover, two Canadian provinces are in the midst of setting up refillable container programs for wine (Leighton 2010). In light of these discussions and initiatives, the alternative packaging scenario incorporates a hypothetical refillable bottle program in Ontario for domestic and bulk imported wines (27.9% of 2008 wine sales by volume – CVA 2008) and spirits (59.9% of spirit sales by volume – ACD 2008). This domestic geographical limitation is selected because, from a logistical standpoint, it does not appear realistic to adopt a system in which used wine and spirit bottles would be returned to their countries of origin for refilling. For packaged wine and spirit imports, the alternative packaging scenario, relative to the 2008 reference, has higher levels of adoption of PET containers for spirits and aseptic cartons for wines. The remaining glass bottles are all replaced with lightweight single use glass bottles. The alternative packaging scenario has market penetration levels of PET containers and aseptic cartons approximately tripled from 2008 levels, for each size of container. However, due to the apparent preference of most wine/spirit consumers for glass containers, I have limited the market penetration of alternative containers to a maximum of 50%, for each container size. These modelling choices result in lightweight container adoption percentages of less than 10% in 2008 increased to 10%. When the 2008 levels of market penetration of PET bottles or aseptic cartons are above 10%, the levels in the alternative packaging scenario are set at 50%. “Other” containers for wines use LSU glass containers while those for spirits are split equally between LSU glass bottles and PET (due to the much greater market penetration of PET packaging for spirits). The use of PET containers to package wine is excluded from the alternative packaging scenario due to the findings of the ISVV (2010) study (see Section 4.1.3) which raises significant doubts about the suitability of such containers for wine. Aseptic cartons are not used for the packaging of spirits in Ontario. There is no agreement in the literature on the mass requirement for a glass container to be considered lightweight. The Waste and Resources Action Programme (WRAP) (2008) report assumed that the lightweight 750 ml single-use glass bottle was 365 g, which is approximately 73% of the mass of the average 750 ml glass wine containers sold in Ontario (see Section 4.3.3). However, since the lowest mass in the sample of 750 ml glass containers measured for this LCA was 393 g, it would seem that 76 the adoption of WRAP’s (2008) assumption might not be appropriate for the Ontario context. Thus, this LCA uses a smaller estimate of a 20% reduction in the average mass of all glass containers. This 20% estimate is also selected in light of the fact that the glass container mass data collected for this study (see Section 4.3.3) do not distinguish between CSU and LSU containers. Since the potential number of refills of lightweight glass bottles would be much lower than for CSU bottles (i.e., 1-5 refills) due to their greater fragility (Saphire 1994), all refillable glass bottles are assumed to have a mass identical to the conventional containers.

4.2.3 System boundary

The LCA system boundary incorporates the following processes (Figure 10): (1) the extraction of raw materials required for each container, and the secondary (i.e., capsules, closures, labels) materials; (2) the transportation of raw materials to processing facilities; (3) raw material processing; (4) the transportation of the processed materials to the container manufacturer; (5) the manufacture of the container; (6) the transportation of the container to the packager; (7) the transportation of the container from the packaging facility to retail outlets in the City of Toronto; (8) the recycling or disposal of the waste packaging materials and (9) the avoided burdens associated with material recycling. Processes within the system boundary also include the production of the capital equipment, infrastructure and energy (fuels and electricity), unless stated otherwise. 77

Primary Packaging Secondary Packaging (containers) (closures, capsules, labels)

Raw material Raw material extraction extraction

Processing Processing

Manufacturing Manufacturing Upstream processes

Packaging

Distribution to retailer Avoided burdens of recycling Bottle washing Use and refilling

Municipal Deposit-Return Downstream collection System processes

Landfilling Recycling

Legend

Environmental emissions Energy and material inputs Material flows between life cycle components

Displaced emissions due to avoided burdens of recycling

Figure 10 System boundary of the LCA of wine/spirit packaging

4.2.4 Data sources

Unpublished data sources for this LCA are provided by the Liquor Control Board of Ontario (LCBO), Tetra Pak Canada, a wine bottle washing equipment manufacturer, as well as wine and spirit producers (Section 4.2.4.2). Field research has been undertaken to estimate the average mass of the various types of containers for wines and spirits sold in Ontario. Other than the LCBO, sources of statistics on wine/spirit consumption and market origin include the Association of Canadian Distillers (2008), the Canadian Vintners Association (2008), and Statistics Canada (2008). Process data associated with the production of the containers and the raw materials from which they are produced are obtained from LCA databases and reports. Those processes taking place in Ontario use the 2008 Ontario electricity production mix, as depicted by Ontario's Independent Electricity System Operator (2009), while those processes undertaken in other countries use the appropriate production mixes derived from International Energy Agency (IEA) 2008 statistics. Since not all of the profile datasets used for this LCA apply identical definitions and assumptions for their parameters, there are instances in which data normalisation is required in order to ensure that all of the data are consistent. Each instance of data normalisation is identified in the appropriate section of this paper.

4.2.4.1 Life cycle assessment software, databases and impact assessment tools

SimaPro 7.2 LCA software is used for the analysis of the net environmental burdens of each LCA scenario. Many of the unit processes incorporated into this LCA use data supplied by the EcoInvent American (US-EI) database, and to a much lesser extent, the Franklin USA 98 database. The ReCiPe v1.02 (Goedkoop et al. 2009) life cycle impact assessment (LCIA) method, with 18 midpoint and 3 endpoint level impact indicators, as well as impact accounting from three modelled perspectives (egalitarian, hierarchical, and individualistic) has been selected to evaluate the environmental impacts of each scenario. ReCiPe depicts LCA impacts in terms of the following endpoint indicators: (1) damage 79

to human health, represented as human health disability-adjusted life year (DALY); (2) damage to ecosystems, expressed as ecosystem species lost per year; and (3) reduction in resource availability, depicted as the surplus cost of resources (Goedkoop et al. 2009). Endpoint impacts are evaluated using ReCiPe’s hierarchical perspective and the “normalization values of the world with the average weighting set” (Goedkoop et al. 2009). The hierarchical perspective is selected for this study since it represents “the most common policy principles with regards to time-frame and other issues” (Goedkoop et al. 2009).

Net greenhouse gas emissions in CO2 equivalents are also calculated and compared using the Intergovernmental Panel on Climate Change (IPCC) 2007 method for 20, 100 and 500 year time horizons.

4.2.4.2 Questionnaire to wine and spirit producers

Data from wineries and distilleries are needed in order to generate estimates of the mean distances that the containers would have been shipped from the manufacturer to the packager (see Section 4.4.3.1). Therefore, in order to acquire this information, I designed and distributed questionnaires to the wineries and distilleries that supplied wines and spirits to the Ontario market in 2008. The 134 questionnaires elicited eleven replies from wineries and distilleries representing approximately 7.2% and 4.0% of wine and spirit sales in Ontario in 2008, respectively. The responses were skewed toward domestic wine producers (94.0% of wine by volume was from domestic suppliers) and foreign spirit producers (98.6% of spirits by volume was from suppliers outside of Canada). Wine company participants were located in North America and Africa, while spirit company participants were North American and European. The low response rates introduce considerable uncertainty to the mean transportation distance estimates derived from the questionnaire responses. This uncertainty is limited to a portion of the packaging transportation component of the life cycle. The sensitivity analysis in Section 4.6 attempts to address this uncertainty by evaluating the life cycle results while using the variable transportation distances (i.e., adding and subtracting the weighted standard deviations). I am unaware of any recent 80

published source of this type of data that could be considered more representative of those circumstances for packaging transportation addressed by this LCA. Additional background information on the questionnaires is available in Appendix 3.

4.2.4.3 Field and laboratory research

Field and laboratory research were needed in order to generate estimates of the average mass of each type of wine and spirit container, closure and capsule supplied to the residents of Toronto. A stratified random sample of 315 full wine and spirit containers (includes the container, its contents, the closure, capsule and label) were weighed at an LCBO outlet using a digital kitchen scale. This sample (165 wine containers and 150 spirit containers, selected among the top 500 most popular products for each type of container sold, by volume of the container) represents 33.1% and 55.0% of 2008 wine and spirit sales in Ontario by volume, respectively. The measured sample was stratified by container size and type to ensure that the measurements represented a minimum of 20% of 2008 LCBO volume sales of wines/spirits in a particular package (e.g., a 750 ml glass bottle). As this field research was undertaken in February and March of 2010, it is possible that the mass of some of the containers might have changed from 2008, although this change is assumed to be negligible. In order to estimate the mass of each empty container, the average mass of the secondary packaging and liquid contents was subtracted from the full container mass measurements. Since the 2008 LCBO volume sales datasets also identified the type of closure used for each product, the mass of the appropriate closure was subtracted from each measured mass value. Paper labels, estimated by WRAP (2008) to weigh 1.9 g, were subtracted from all measurements except from those identified as aseptic cartons. The mean mass value of the average capsule (1.4 g) was only subtracted from those measurements associated with products that had corks. The mass of wine subtracted from each measurement was based upon the average density of wine, estimated at 1g/ml (WRAP 2008), and the volume of the particular container. In general, wine densities vary relative to their alcohol and sugar content. However, the variation is relatively small. For example, in their sample of 58 red and white wines, Bavcar and Kosmerl 81

(2002) measured the range in average densities to be 0.989-1.028 g/L (conversion from relative density (specific gravity) to average density by author). For spirits, the average densities are more variable, and must be taken into account in order to generate accurate estimates of the mass of each type of container. As I was unable to locate scholarly sources for spirit densities, laboratory measurements were undertaken to determine the densities of 25 spirit brands. The mass of 50 ml of each spirit, measured in a 100 ml graduated cylinder, was measured using a Sartorius Research balance with a measurement readability of 10-5 g and a measurement standard deviation less than or equal to ± 0.05 mg. This balance was also used to weigh closures and capsules. Measured spirit densities ranged from 0.939-1.136 g/L. The highest spirit density was 21% greater than the lowest, a much larger difference than for the wine densities measured in Bavcar and Kosmerl (2002), which had a 4% difference. To verify the accuracy of the measured spirit densities, the estimated mass of the largest PET container (1750 ml), based on the procedure described previously, was compared with the measured masses of empty PET spirit containers of the same volume. The original estimate of 87 g (derived from the measured masses of full containers from 11 brands) compared quite favourably with the mass measurements of two empty 1750 ml containers (including labels) for spirits, with one having a mass of 87 g, and the other with a mass of 89 g. The detailed results of the field and laboratory work are listed in Appendix 3.

4.3 LCA input profile

Liquor Control Board of Ontario (LCBO) volume sales datasets for wines/spirits for the year 2008 are used to calculate the percentages of wines/spirits packaged in each type and size of container, and the percentages using each type of closure. Wine and spirit sales at the LCBO do not represent all sales of these products in the province, although they do represent the majority (83.6% of Ontario’s wine volume sales in 2007- 2008, and almost all of the provincial market for spirits – LCBO 2008). Therefore, the LCBO data were normalized to the 2008 “over the counter” volume sales figures for Ontario published by the Association of Canadian Distillers and the Canadian Vintners 82

Association. The resulting figures were then multiplied by 19.76%, the percentage of the province’s population living in the City of Toronto in 2006 (City of Toronto 2006, Statistics Canada 2006). The alternative wine and spirit packaging options do not pertain to ciders and coolers, the latter of which blend wines/spirits with non-alcoholic beverages and possess an alcohol content of between 5% and 10%. Therefore, volume sales data for coolers and ciders needed to be removed (see Appendix 3 for procedure). As described in Section 4.2.4.3 on field and laboratory research, the masses of a large sample of different containers are used to provide estimates of the average mass of each type of container. Figure 11 depicts the results from this work, which reveal the extent of the differences in the mass of glass, PET and aseptic containers, by volume (see Appendix 3 for additional information).

1400

1200

1000

800

600

Mass of container (g) 400

200

l l l l l l l m m ml ml ml m ml m 0 ml 0 0 0 ml 5 m 0 m 0 0 m 5 ml 0 ml 0 0 ml 5 7 5 0 7 5 5 7 00 50 00 200 ml 3 7 140 750 75 3 7 14 000 1 1 T T 1 1 -Gl Gl -Gl -Gl ET T C -Gl 2 S- S S Gl 1 Gl 1 P PE PE A W S- S- PE W-Gl W-Gl W W-Gl 3000 ml W- S-PET 200S- ml S- W- W-AC 1500 W-PET 10 S- S-PET 17 Container type (wine/spirits, material type, size by volume)

Figure 11 Mass of wine and spirit containers by container type. Acronyms: W-Gl: glass container for wine; S-Gl: glass container for spirit; W-PET: PET container for wine; S-PET: PET container for spirit; and W-AC: wine aseptic carton. The “whiskers” (the vertical lines drawn from the top edges of the boxes depicting the container masses) indicate the largest and smallest values within the sample. Boxes without whiskers either have a sample size of one, or use an alternative method to determine the average mass of the container. More detailed information on the procedure to determine the average container masses is available in Appendix 3.

The LCBO 2008 volume sales datasets are also used to determine the percentages of wine and spirit containers using various types of closures (corks, synthetic corks, metal and plastic screw caps, pressure tabs, and others). LCA literature (e.g., PwC 2008) and my own measurements provide the mean closure mass and composition data required. The results of the 2010 Wine Capsule Survey by Wine Business Monthly (Fisher 2010) are used to provide estimates of the popularity of the various types of capsules used to cover the tops of those glass containers using corks. According to the American survey, approximately 67% of wine capsules for wines between US$7-13.99 were polylaminates (a composite of aluminum and polyethylene), 21% were composed of polyvinylchloride (PVC), 8% were composed of tin and 6% were aluminum (see Appendix 3 for calculation). Although the LCBO 2008 volume sales data indicates that the presence of capsules on spirit containers is far less common than on wine bottles, capsules on spirit bottles were also taken into account in this LCA. The composition of the secondary packaging changes with the type of primary container used. For example, if an aseptic carton is substituted for a glass container, the aluminum screw cap or cork would be replaced with a plastic cap. Appendix 3 lists the estimated mass, material composition and adoption levels of both capsules and closures, by type, as well as the procedures used to generate these estimates. Table 6 depicts the mass of wine and spirit packaging in the 2008 reference and alternative packaging scenarios. The alternative scenario has a 46.9% reduction in the mass of packaging waste in comparison to the 2008 reference, equal to 9,248 tonnes. Losses during manufacturing and packaging were taken into account for containers, but considered negligible for secondary packaging.

Table 6 Mass of wine and spirit packaging material inputs and waste under the 2008 reference and alternative packaging scenarios

Wine/spirit packaging mass (tonnes) Containers Secondary packaging Total (mass of inputs and waste are identical) Aseptic Glass PET Closures Capsules Labels cartons 2008 Inputs: Inputs: Inputs: 169 21 70 Inputs: reference 18 19,753 91 20,122 Waste: Waste: Waste: Waste: 17 19,358 90 19,724 84

Alternative Inputs: Inputs: Inputs: 166 20 70 Inputs: packaging 35 10,255 141 10,687 Waste: Waste Waste: Waste 33 10,050 138 10,476

4.4 LCA unit processes

With some exceptions, unit processes from the US EcoInvent (US-EI) and Franklin USA 1998 databases are used for the LCA scenarios. Unit processes are also defined by the author for (1) aseptic carton manufacture; (2) bottle washing; (3) sorting of recyclables at material recovery facilities; and (4) PET recycling; with data acquired from the published scientific literature, equipment manufacturers and operators. These author- defined unit processes are produced when the available processes are inadequate. A comprehensive list of unit processes incorporated into the LCA scenarios, and the quantities required of each process, are listed in Appendix 3. For those unit processes depicting the production of container packaging materials, the default supply mixes for their electricity inputs are replaced with the average electricity supply mix from the 2008 national electricity transmission grid of the country from where the imported wines and spirits originated. For domestically-sourced wines and spirits, the 2008 Ontario electricity supply mix is used in the calculation of the environmental emissions resulting from the production of the containers. The compositions of each electricity supply mix are acquired from International Energy Agency (IEA) 2008 statistics and Ontario's Independent Electricity System Operator (2009), and listed in Appendix 3.

4.4.1 Aseptic carton production

An aseptic carton is composed of paperboard, polyethylene and aluminum (Tetra Pak 2005). Unlike most other wine and spirit containers, aseptic cartons lack a paper label. One litre aseptic cartons sold in Toronto are almost exclusively tetra prismas, whereas tetra briks are used for 1.5 litre containers. Smaller sizes of aseptic cartons for wine are also available, but quite uncommon, and accounted for far less than 1% of wine volume sales in 2008 (LCBO 2008). 85

Since one litre aseptic cartons for wine originate from Lindberg, Germany (J. Koel - Tetra Pak Canada, pers. comm. 2008), the average European electricity supply mix depicted in the EcoInvent database is selected for those processes associated with aseptic cartons, namely primary aluminum, solid bleached board, and LDPE packaging film. As there is no process provided for aseptic carton manufacture in the given database, one is defined based upon a dataset published by Tetra Pak Inc. (2003). These data indicate that material loss at the manufacturing facility amounts to 5.2%, with almost all of this waste material recycled (99.4%), and the remainder incinerated. This level of material loss is incorporated into the aseptic carton manufacturing process defined by the author in SimaPro. Environmental burdens associated with the production of the inks for the containers are excluded from this LCA due to a lack of available data.

4.4.2 Glass bottle production

The unit processes for glass manufacturing encompass the extraction of raw materials, their transportation to the glass manufacturing facility, batching (mixing primary materials), melting and refining the materials, and fabrication of the glass bottles themselves. Since brown, clear, and green glass containers are used to package wines and spirits, and the means of producing each of these types of glass differ somewhat, it is necessary to determine the proportions of wine/spirit bottles of each colour. The colours of the container glass were identified for the top one hundred glass wine containers sold by the LCBO in 2008, stratified by container size and product brand. The results of the sample indicated that 29.5% of the containers were clear, 69.3% were green, and 1.2% were brown (sample size represented 32.1% of 2008 LCBO wine container sales). For the top one hundred spirit containers, 78.0% were clear, 17.6% were brown, and 4.4% were green (sample size represented 47.4% of 2008 LCBO spirit container sales). These percentages are assumed to be representative of the wine and spirit container sales in Toronto in 2008 and are used as the percentages of each type of glass included within the LCA scenarios. The proportion of recycled cullet in the glass manufacturing process affects its material and energy inputs, as well as its environmental emissions. The recycled content 86

figures for brown (55%), clear (58%) and green (75%) glass are taken from Magaud et al. (2010) for Canadian glass containers and incorporated into both LCA scenarios. As the glass production processes within the EcoInvent database do not incorporate the same percentages of recycled content (glass cullet) as those derived for the Toronto context, it was necessary to alter them. These alterations are based on the method proposed by Magaud et al. (2010), who adapted the EcoInvent unit processes for glass used in Quebec (Canada) beer containers (see Appendix 3 for additional details).

4.4.3 Polyethylene terephthalate bottle production

Unit processes associated with PET bottles include granulate bottle grade PET and PET bottle forming through preform injection molding, followed by blow molding. The recycled content of the PET containers is considered negligible, based on the 2008 statistics indicating that 2.4 million tonnes of PET bottles were supplied to the United States, while 89 000 tonnes were used as inputs for PET container production (NAPCOR 2008).

4.4.4 Production of secondary packaging

Packaging systems for wines and spirits include not only the containers, but also the closures, labels and capsules. The main types of closures for wine and spirit containers are aluminum screw caps, corks, and synthetic closures composed of polyurethane or styrene-butadiene-styrene (SBS) polymers (LCBO 2008). Capsules are used for containers that have closures other than screw caps, plastic and metal lids. The life cycle of the materials used for the secondary packaging are depicted by US-EI unit processes, except for corks, which use the EcoInvent European unit process since corks are produced in Europe.

4.4.5 Transportation of materials

Calculating the environmental emissions from the transport of containers to market requires information on the distances travelled and the means of transportation. Data on the average number of wine and spirit containers carried in each shipping container and the proportion of vehicles used for backhaul are not necessary due to the method of accounting for transport in the US-EI database. Tonne-kilometres (tkm) are used to depict “the transport of one tonne of goods by a certain means of transportation over one kilometre” (Spielmann et al. 2007). The selected unit process to depict transportation by truck is “transport, lorry>28 tonnes, fleet average” which uses the same type of truck as that used by Magaud et al. (2010) for the transportation of various types of beer containers in Canada. This unit process, as well as the unit processes for transportation by ship and rail, adopts the average utilization of loading capacity of each freight transport mode into the process design. The transportation emissions from the closures, capsules and labels are deemed negligible due to their low masses relative to the container, and therefore are excluded from the LCA. Table 7 displays the estimated average distances that wine/spirit containers travelled to Toronto in 2008. The procedures used to estimate these figures are described in Sections 4.4.5.1 and 4.4.5.2, as well as Appendix 3.

Table 7 Average distances travelled by wine/spirit containers from the manufacturer to the packager and from the packager to Toronto in 2008

Average distances travelled, by mode of transportation (km) SHIP TRUCK RAIL From manufacturer to packager 6.11*102 4.12*102 4.12*102 From packager to Toronto 9.50*103 7.32*102 5.57*102 Total upstream transportation distance 1.01*104 1.14*103 9.69*102

4.4.5.1 Transportation of containers from manufacturer to packager

There is a relative dearth of estimates of the mean distance that empty containers are transported to wine and spirit packaging facilities. Franklin Associates (2006) assumes a mean transport distance of 154 km, based on the particular circumstances for wine bottling in Northern California. This estimate differs substantially from the estimate of a 426 km mean distance for glass transportation from manufacturer to market 88

that was used by ICF Consulting (2005) in its LCA of GHGs from waste management in Canada. However, the ICF Consulting estimate uses the transportation of steel as a surrogate for glass. In order to acquire more representative transport data, questionnaires were sent to firms supplying wines/spirits to the Ontario market. Responses to my questionnaires were received from eleven wine/spirit firms supplying approximately 10.7 million litres to the Ontario market. These data indicate a weighted mean transport distance of 611 km by ship and 824 km by land (see Appendix 3 for calculation procedure).

4.4.5.2 Transportation of containers from packager to Toronto

Statistics Canada (2008) data on packaged wine and spirit imports to Ontario in 2008 are used to identify the countries of origin of these products. Wines and spirits not shipped in bulk are assumed to be packaged in their countries of origin. Bulk imported wines and spirits packaged in Ontario are considered "domestic" in the Statistics Canada, LCBO, Canadian Vintners Association (CVA) and the Association of Canadian Distillers (ACD) statistics (e.g., wines bearing the designation "Cellared in Canada" are domestic). Only those transportation emissions attributed to the wine/spirit container are taken into account, while those emissions resulting from the container contents are outside of the system boundary because the focus of the LCA scenarios is the packaging itself. Statistics Canada data on the volume of wines and spirits imported to Ontario are used to identify the quantities of wine/spirit imports from each country concerned (Figures 12 and 13). These data differ to a small extent from the total import estimates of the ACD (2008) and CVA (2008), which do not break down the import volume data by country of origin. Therefore, the amounts of packaged wine and spirit imports by country of origin from Statistics Canada (2008) are normalized to the total import sales estimates from the ACD and CVA by multiplying the Statistics Canada wine import figures by 1.041 and the spirit import figures by 0.933. The resultant figures are then multiplied by the percentage of the population of Ontario located in the City of Toronto in 2006 (19.76%) (City of Toronto 2007, Statistics Canada 2006) in order to provide approximations of the volumes of wines and spirits that were consumed by the residents 89 of Toronto in 2008. The distances between international ports are obtained from the World-Ports Distances Calculator (2010). For distances traveled using land-based transport, the first distance supplied through Google Maps (Google 2010) is selected.

New Zealand Germany 2% 2% Other 3% Spain Portugal 3% 3% Italy 24% South Africa 4% Argentina 4%

Chile 9%

Australia 18% US-California 12%

France 16%

Figure 12. Country / US state of origin of imported packaged wines (% of volume imported to Ontario in 2008). Percentages based upon Statistics Canada (2008) data.

Poland 2% Other Finland 8% 2% Cuba United Kingdom 2% 24% US-Tennessee 2% US-Florida 3% Mexico 3% US-Illinois 4% US-California Russian Federation 4% 13% Ireland, Republic of (EIRE) 4% US-Kentucky 4% Italy France 5% Sweden 12% 8%

Figure 13. Country / US state of origin of imported packaged spirits (% of volume imported to Ontario in 2008). Percentages based upon Statistics Canada (2008) data. 90

4.4.6 Rinsing and filling of containers

Once shipped to the packaging facility on pallets, glass and PET bottles are rinsed/sterilized, filled, corked/capped, labelled, boxed, palletized and shrink-wrapped (Mata and Costa 2001, Rockwell Automation Inc. 2010). Aseptic cartons, which are shipped as rolls, are sterilized, opened, filled, capped and packaged for shipping (Mourad et al. 2008). The Franklin Associates (2006) LCA of wine packaging deemed the differences between the packaging scenarios for the rinsing/washing and filling components of the life cycles to be negligible, and omitted these components. Due to a lack of recent available data pertaining to these components of the life cycle, the same assumption of the negligibility of the differences applies. However, there is significant loss of material at packaging facilities. WRAP (2008) estimates a 2% loss during glass and PET wine bottle filling. This container loss estimate is adopted for all types of wine and spirit containers in the LCA scenarios.

4.4.7 Reuse of containers

For the alternative packaging and RFG scenarios, it is assumed that empty refillable wine and spirit containers are brought to The Beer Store, which administers the Ontario deposit-return program. Each refillable glass bottle would be filled 15 times (14 refills), an assumption identical to that adopted by The Beer Store, which operates the deposit-return system for alcoholic beverages in Ontario (TBS 2009). The number of times each bottle can be refilled, or the trippage rate, is dependent upon the mass, strength and condition of the bottle, which affects the breakage rate, cited by industry sources to be between 3% and 5% (Saphire 1994). The trippage rate is also limited by the rate of container return. Limiting the configuration of refillable containers to a few types of industry standard bottles would facilitate the operation of the refilling system, and would in all probability be implemented should refilling on a large-scale occur.

4.4.7.1 Transportation of used bottles between retailer and cleaning/refilling facility

The two-way transport distance of empty wine/spirit glass bottles between the retailer in Toronto and the bottlewashing/refilling facilities is assumed to be 200 km (i.e., approximately the round-trip distance between Toronto and the Niagara wine region between Stoney Creek and Niagara on the Lake). This appears reasonable in light of the suggestion by Valiante (2007) that the typical round trip for a refillable wine bottle in Ontario “might be in the order of just 300 kms.” As this case study focuses on the City of Toronto, the largest market in Ontario, it is likely that the distance would be substantially less than Valiante’s (2007) estimate. Once washed and refilled, the containers would be transported back to Toronto.

4.4.7.2 Washing of used glass bottles

The cleaning and refilling of glass bottles requires caustic soda, warm/hot water, fuel and electricity. Since LCA processes for bottle washing depicted in LCA databases tend to be out of date (e.g., the BUWAL 250 LCA process entitled “washing bottles” uses data from the early 1990s), it was necessary to obtain recent technical information. Confidential bottle washing data for modern equipment used to clean 1000 ml glass wine containers were recently provided to the author by a bottle washing equipment manufacturer in order to facilitate the design of an LCA process within SimaPro 7.2. These data are used as inputs for the bottle washing unit process designed for the alternative packaging and RFG scenarios. However, due to a lack of sufficient information, the environmental burdens resulting from the manufacture of the bottle washing machine and the construction of the facility are excluded from the defined process.

4.4.8 Waste management

The waste management stage of the life cycle includes waste collection, transfer, sorting, recycling, and disposal. The wine/spirit containers are collected through two 92 programs – the residential waste management system of the City of Toronto, and the “Bag-it-back” deposit-return program for wine and spirit containers, which has been operating in Ontario since 2007 (TBS 2009). The container return rates used are the same as those during the 2007-2008 fiscal year of the deposit-return program: 29% for aseptic cartons, 34% for PET bottles, and 69% for glass bottles (TBS 2009). Those containers collected through the municipal program are assumed to be recycled at levels identical to the City of Toronto’s 2008 municipal recycling rates for each respective waste stream (glass, PET, and poly-coat (for aseptic cartons)). The secondary packaging materials, other than the paper labels for the containers retuned through the deposit-return program, are landfilled.

4.4.8.1 Waste collection, transportation and sorting

Once used, wine and spirit packaging can be allocated to the deposit-return system, the municipal recycling system, or the municipal landfill. For the first option, the environmental burdens which would result from transporting used wine and spirit containers from household to collection point (The Beer Store) are considered negligible relative to those resulting from shipment to the recycling locations. The assumption of negligibility may be supported by the possibility that: (1) some households may use non- vehicular means of transport – walking and biking – to return the containers; and (2) the trips to The Beer Store retailers could also have multiple objectives, with the obvious example of purchasing beer. Outside of the deposit-return system, discarded wine and spirit packaging materials may also be collected at the curbside by municipal waste collection trucks and either sent to a material recovery facility (MRF) for sorting, or transferred to larger vehicles for transport to the landfill. Estimates of the average distance (19.75 km) that Toronto’s waste collection vehicles transported waste, as well as the locations of the glass, PET and aseptic carton recycling facilities, are based on unpublished datasets provided to the author by the City of Toronto Solid Waste Management Services (2010a) for this LCA. 93

Municipal recyclables, which are collected in a commingled state, are sorted before sending the material onward to the recycling processors. Waste allocated to landfill (garbage) is not sorted. Appendix 3 provides additional detail on the data collected and methods applied for estimating the environmental impacts from waste collection, sorting and transportation.

4.4.8.2 Recycling

Avoided burdens are taken into account when depicting recycling. However, the material inputs for the recycling processes are often not equal to the quantities of recycled material produced due to processing losses and the presence of contaminants. Thus, the avoided burdens of recycling are based on a smaller quantity of material than the recyclable material inputs. Although some glass is not recycled back into its original product type (i.e., a glass container), lack of sufficient published data on the recycling of glass into fibreglass and other products necessitated the assumption of a closed-loop “bottle-to-bottle” recycling system for the City of Toronto’s waste glass. Unlike municipal collection, the Province of Ontario’s deposit-return system for wine and spirit containers allows glass to be separated according to colour, and recycled into glass bottles. Non-refillable clear glass bottles collected through this system are remanufactured into new bottles, while 50% of coloured glass bottles are recycled into fibreglass, with the remainder used for manufacturing new coloured containers (TBS 2009). Forms of open-loop recycling for glass in Toronto may include using glass as an aggregate substitute, a sand substitute, an abrasive material, a filter material, and as a raw material in the fibreglass industry (Unical 2008, Toronto City Council 2002). Although recycled PET bottles are frequently used as inputs for plastic strapping and felted automotive materials (TBS 2009), this LCA assumes a “bottle-to-bottle” recycling system. In comparison, the recycling of aseptic cartons is more complicated due to their nature as a composite package. In the recycling process, these cartons are hydro-pulped, permitting the cardboard component to be used as pulp, with the remaining aluminum and plastic landfilled (CM Consulting 2010). This aseptic carton recycling 94

process is incorporated into the LCA. Since 2010, aseptic cartons can be recycled through thermokinetic mixing, which combines the three materials to generate a product suitable as inputs for the production of goods such as flower pots and plastic lumber (CM Consulting 2010).

4.4.8.3 Landfilling

The US-EI landfill unit process is employed to calculate the environmental emissions from those packaging materials that are landfilled. In 2008, approximately 93.8% of the City of Toronto’s residential waste was sent to Michigan (441 km), with the remainder sent to the Green Lane landfill located between St. Thomas and London, Ontario (203 km) (City of Toronto 2010a).

4.5 Life cycle impact assessment (LCIA) results

The LCIA results are compared using the endpoint level indicators of ReCiPe, although the midpoint level impact results are also available in Appendix 3. ReCiPe’s damage to human health indicator is measured using disability adjusted life years (DALYs), equal to the sum of years of life lost and years of life disabled, the latter of which using a subjective valuation of health disabilities on quality of life (Goedkoop et al. 2009). The characterisation of this endpoint indicator includes the following midpoint level impacts: climate change, particulate matter formation, human toxicity, ionizing radiation, photochemical oxidant formation and ozone depletion. The damage to ecosystem diversity endpoint indicator assumes that species diversity is a suitable indicator of ecosystem quality. This indicator is measured in terms of the potentially disappeared fraction of species (PDF) integrated over area (or volume, in aquatic and marine habitats) and time (Goedkoop et al. 2009). Although all species are assumed to be of equal importance, the default species densities are far higher in terrestrial environments than in freshwater and marine ones. The characterisation of the damage to ecosystem diversity endpoint indicator includes the following midpoint level impacts: climate change, agricultural land occupation, natural land transformation, urban 95

land occupation, terrestrial acidification, terrestrial ecotoxicity, aquatic eutrophication, aquatic ecotoxicity, marine ecotoxicity, and marine eutrophication. The damage to resource availability indicator quantifies its impact as a marginal increase in resource extraction costs due to the resource consumption attributed to the life cycle (Goedkoop et al. 2009). The characterisation of this endpoint level indicator includes the mineral resource depletion and fossil fuel depletion midpoint level impacts.

4.5.1 Individual package scenarios

The comparative results for the individual package LCA scenarios indicate that the refillable glass containers and aseptic cartons (for wine only) are responsible for the lowest net endpoint level impacts, followed by PET containers, lightweight single use glass bottles, and conventional single use glass bottles. The reduction in life cycle impacts from replacing CSU glass containers with LSU glass bottles is approximately equal to the proportion of the mass reduction from the substitution, for all three of the endpoint level impacts. In comparison, the net life cycle impacts from CSU and LSU bottles are almost always more than double the magnitudes of those of the remaining container types. Endpoint level impacts for wine containers are almost identical for refillable bottles and aseptic cartons, with the exception of the damage to ecosystem quality impact, in which the RFG bottle has half the potential impact of the AC. Aseptic cartons are responsible for greater ecosystem quality damage because of the land use required to supply the wood pulp for the paperboard input, an input not required for refillable bottles. Table 8 depicts the net endpoint level impacts from the life cycles of the alternative packaging systems as percentages of the net impacts generated by the CSU glass packaging system.

Table 8 Net endpoint level impacts from the life cycles of the alternative packaging systems as percentages of the net impacts of the CSU glass packaging system

Endpoint level CSU glass % of CSU packaging impact impact packaging AC packaging LSU packaging PET packaging RFG packaging impact impact impact impact impact 1 L wine packages Ecosystem 8.21*10-9 33.0% 84.5% 44.1% 15.5% diversity species*yr Human health 1.64*10-6 DALY 16.5% 85.9% 41.5% 17.1% Natural resources $3.19 19.2% 86.8% 67.4% 19.2% 96

750 ml spirit packages Ecosystem 7.81*10-9 N/A 80.8% 44.1% 13.2% diversity species*yr Human health 1.57*10-6 DALY N/A 80.8% 39.2% 13.6% Natural resources $3.21 N/A 81.0% 66.% 15.5%

The results for the three endpoint indicators tend to be similar in terms of the relative magnitudes of each of the packaging life cycles (Figures 14, 15 and 16). For all containers except refillables, container production contributes more than half of the absolute life cycle impacts. The relative avoided burdens of recycling vary substantially using each indicator. For example, the avoided burdens from PET recycling are similar to those from recycling under the CSU scenario, except for the resource depletion indicator in which the net benefit of recycling is much greater than that of recycling in the CSU packaging life cycle. In this case, the change is due to the design of the resource depletion indicator, which accounts for fossil fuel and metal depletion in the evaluated scenarios. For PET recycling, there are not only savings in fuel inputs, as in glass recycling, but also in the fossil fuels no longer required as material inputs for PET. Both the aseptic carton and PET bottle show far larger endpoint impacts from container production than do the glass containers, per unit of container mass. However, the overall decrease in container mass from the AC and PET substitutions, relative to the volumes of wines/spirits that can be packaged, more than compensates for the increased impacts per unit of container mass. The small masses of the AC, PET and RFG containers relative to the CSU and LSU glass bottles dramatically reduce the impacts from transportation. The impacts from the non-reusable nor recycled secondary packaging are similar for all wine containers but the ACs, which use plastic screw caps that generate far smaller potential impacts. However, the secondary packaging impacts for spirit containers are consistently lower than for wine containers due to the higher level of adoption of plastic screw caps instead of metal ones. The relative importance of secondary packaging within the life cycles is most important for RFG bottles for wine, with this component responsible for greater impacts than the production of the reusable glass bottle (impacts of primary package are divided by 15 in order to reflect the burdens when using the one litre and 750 ml functional units). The water depletion midpoint level 97 impact, which is omitted from all of the endpoint level indicators, is lowest for RFG containers. The primary differences in the inputs for the wine and spirit package LCAs include the masses of the containers, the proportions of the glass containers that are clear, green and brown in colour, and the composition of the secondary packaging.

1.00E-08

8.00E-09 Refilling

Landfilling 6.00E-09 Waste collection and transportation

4.00E-09 Packaging transportation Secondary packaging 2.00E-09 Primary packaging

Ecosystem speciesyear lost per Recycling 0.00E+00

-2.00E-09 1 L AC 1 L 1 L 1 L 1 L 750 ml 750 ml 750 ml 750 ml CSU LSU PET RFG CSU LSU PET RFG

Figure 14 Impact values for the damage to ecosystem diversity endpoint level indicator, identifying the contribution of each LCA component for individual one litre wine and 750 ml spirit packages. AC – Aseptic carton. CSU – Conventional single use glass bottle. LSU – Lightweight single use glass bottle. PET – Polyethylene terephthalate bottle. RFG – Refillable glass bottle. The refilling component includes both bottle washing and transportation to and from the location of refilling. 98

2.00E-06

1.50E-06 Refilling

Landfilling

1.00E-06 Waste collection and transportation Packaging transportation Secondary 5.00E-07 packaging Primary packaging

Recycling Disability adjusted life year (DALY) year life adjusted Disability 0.00E+00

-5.00E-07 1 L AC 1 L 1 L 1 L 1 L 750 ml 750 ml 750 ml 750 ml CSU LSU PET RFG CSU LSU PET RFG

Figure 15 Impact values for the human health endpoint level indicator, identifying the contribution of each LCA component for individual one litre wine and 750 ml spirit packages.

3.5

3 Refilling

2.5 Landfilling

2 Waste collection and transportation

1.5 Packaging transportation Secondary 1 packaging Primary packaging 0.5 Surplus cost of resources ($) resources of cost Surplus Recycling 0

-0.5

-1 1 L AC 1 L 1 L LSU 1 L PET 1 L 750 ml 750 ml 750 ml 750 ml CSU RFG CSU LSU PET RFG

Figure 16 Impact values for the resource depletion endpoint level indicator, identifying the contribution of each LCA component for individual one litre wine and 750 ml spirit packages. 99

4.5.2 Municipal scale scenarios

Endpoint level results for the 2008 reference packaging scenario evaluated using the ReCiPe v1.02 hierarchical method indicated that the production of the container is responsible for between 57% and 68% of the absolute life cycle impacts. The second most influential impact over the life cycle is that of container transportation (16%-23%). The significance of secondary packaging (4%-5%) is slightly less than half that of recycling (9%-12%). Landfilling is negligible under both scenarios since the majority of the waste is glass – an inert material. The refilling component of the alternative packaging scenario life cycle, which includes bottle washing and transportation to and from the bottle washing facilities, is also negligible. For all impact categories, the recycling component is responsible for the net avoided burdens, while the remaining components are responsible for the environmental damage. The alternative packaging scenario reduces the significance of the impact of the primary package and upstream transportation relative to the other LCA components, especially that of secondary packaging. The human health impact falls by 42%, while the impacts on ecosystem quality and the scarcity of resources are reduced by 42% and 40%, respectively. Glass production, fossil fuel production and use, and transportation are the largest process contributors to the three endpoint level impacts.

4.5.2.1 Climate change

The IPCC 2007 method is used to calculate the global warming potential (GWP) of each scenario using 20, 100 and 500 year time horizons (Figure 17). The GWP from 4 4 the 2008 reference scenario ranged from 2.71*10 to 2.98*10 tonnes of CO2 equivalents, while that of the alternative packaging scenario ranged from 1.56*104 to 1.74*104 tonnes. The adoption of the 20 year time horizon produced somewhat larger estimates than for the longer time horizons. The top processes contributing to these emissions are associated with the production and processing of natural gas, which is used extensively in glass and plastics production. The results from the climate change damage indictors of ReCiPe indicate that primary packaging, the greatest contributor to GWP in both 100 scenarios, is responsible for approximately three times the climate change forcing of packaging transportation, the second most important contributor. The reduced production of container materials and tonnage of packaging for transportation causes the greatest decreases in the GWP impact.

20 yr 100 yr 500 yr 3.5E+07

3.0E+07

2.5E+07 Refilling

equivalents) Landfilling 2 2.0E+07 Waste collection and transportation

1.5E+07 Packaging transportation Secondary packaging (closures, capsules, labels) 1.0E+07 Primary packaging (glass, PET, aseptic cartons) Recycling (avoided burden) 5.0E+06

Global warming potential (kg CO 0.0E+00

-5.0E+06 Ref Alt Ref Alt Ref Alt

Figure 17 Impact values for global warming potential using 20, 100 and 500 year time horizons and identifying the contribution of each LCA component for the municipal scale scenarios (IPCC 2007 calculation method employed).

4.5.2.2 Endpoint level impacts

ReCiPe’s endpoint level impacts include damage to ecosystem diversity, damage to human health, and damage to resource availability. The results for all three impact types are very similar in terms of the relative contributions of the life cycle stages to the damage impacts of the 2008 reference and alternative packaging scenarios (Figures 18, 19 and 20). The similarity is based upon the importance of fossil fuel use, and its climate change impact, on the endpoint level impacts, reaching between 66% and 99% of the impacts (Figures 21, 22, and 23). 101

0.4

0.35

0.3 Refilling

0.25 Landfilling

Waste collection and transportation 0.2 Packaging transportation

0.15 Secondary packaging (closures, capsules, labels) Primary packaging (glass, PET, 0.1 aseptic cartons) Recycling (avoided burden)

Ecosystem species lost per year per lost species Ecosystem 0.05

-0.05 2008 reference scenario Alternative packaging scenario

Figure 18 Impact values for the damage to ecosystem diversity endpoint level indicator, identifying the contribution of each LCA component for the municipal scale scenarios.

60 Refilling

50 Landfilling

Waste collection and transportation 40 Packaging transportation

30 Secondary packaging (closures, capsules, labels) Primary packaging (glass, PET, 20 aseptic cartons) Recycling (avoided burden)

Disability-adjusted lifeyear (DALY) 10

-10 2008 reference scenario Alternative packaging scenario

Figure 19 Impact values for the damage to human health endpoint level indicator, identifying the contribution of each LCA component for the municipal scale scenarios. 102

1.75E+08

1.50E+08

1.25E+08 Refilling

Landfilling 1.00E+08 Waste collection and transportation

7.50E+07 Packaging transportation Secondary packaging (closures, capsules, labels) 5.00E+07 Primary packaging (glass, PET, aseptic cartons) Recycling (avoided burden) Surplus cost of resources ($) 2.50E+07

0.00E+00

-2.50E+07 2008 reference scenario Alternative packaging scenario

Figure 20 Impact values for the damage to resource availability endpoint level indicator, identifying the contribution of each LCA component for the municipal scale scenarios.

For both the 2008 reference and alternative packaging scenarios, most of the impact on ecosystem diversity results from climate change, as well as agricultural and natural land occupation (Figure 21). The most important process contributions are natural gas combustion, glass production and forest harvesting to produce the cardboard for the aseptic cartons.

103

2008 Reference

Climate change Agricultural land occupation Natural land transformation Urban land occupation Terrestrial acidification Terrestrial ecotoxicity Freshwater eutrophication Waste Prevention Scenario

0% 20% 40% 60% 80% 100% % of the sum of the Damage to Ecosystem Diversity endpoint impact values

Figure 21 Percentage contributions of midpoint level impacts to the damage to ecosystem diversity endpoint level damage indicator values, for the municipal scale scenarios.

The climate change component is by far the greatest contributor to ReCiPe’s damage to human heath indicator, at more than 60% of the net impact, under both scenarios (Figure 22). Particulate matter, mainly from upstream transportation, is responsible for 30% of the impact. Glass production, natural gas combustion and transport by ship and truck cause most of the human health damage from wine and spirit packaging.

104

2008 Reference

Climate change Particulate matter formation Human toxicity Ionising radiation Photochemical oxidant formation Ozone depletion Waste Prevention Scenario

0% 20% 40% 60% 80% 100% % of the sum of the Damage to Human Health endpoint impact values

Figure 22 Percentage contributions of midpoint level impacts to the damage to human health endpoint level damage indicator values, for the municipal scale scenarios.

Fossil fuel consumption is responsible for 99% of the endpoint level impact for damage to resource availability, with the remaining 1% resulting from the mining of metals used in the secondary packaging (Figure 23). The alternative packaging scenario reduces fossil fuel consumption by 40% whereas metal mining falls by only 10% because the secondary packaging is not subject to lightweighting.

105

2008 Reference

Fossil depletion Metal depletion

Waste Prevention Scenario

0% 20% 40% 60% 80% 100% % of the sum of the Resource Availability endpoint impact values

Figure 23 Percentage contributions of midpoint level impacts to the damage to resource availability endpoint level damage indicator values, for the municipal scale scenarios.

4.6 Sensitivity analysis

The results displayed in Figures 14 to 20 indicate that the transportation of packaging is a significant contributor to the life cycle impacts of all of the LCA scenarios. This significance, along with the considerable uncertainty of the container transportation distance estimates (i.e., 611±1765 km by ship, 412±405 km by train, and 412±405 km by truck) derived from replies to the questionnaires sent to wine and spirit companies, makes packaging transportation an appropriate component to examine using a sensitivity analysis. For both the municipal and individual package LCAs, the sensitivity analysis addresses the impacts from transporting containers from manufacturer to packager. The weighted standard deviations of the container distance estimates are used as a guide for the design of the sensitivity analysis. For transportation by ship, the distance is changed from 611 km to 0 km and 2376 km, while the distances by truck and train are changed to 7 km and 817 km. 106

Table 9 Values used in the sensitivity analysis to depict the distances travelled by wine/spirit containers from the manufacturer to the packager and from the packager to Toronto in 2008

Average distances travelled, by mode of transportation (km) SHIP TRUCK RAIL Distance based on subtraction of weighted standard deviations From manufacturer to packager 0 7 7 From packager to Toronto 9.50*103 7.32*102 5.57*102 Total upstream transportation distance 9.50*103 7.39*102 5.84*102 Distance based on addition of weighted standard deviations From manufacturer to packager 2.38*103 8.17*102 8.17*102 From packager to Toronto 9.50*103 7.32*102 5.57*102 Total upstream transportation distance 1.19*104 1.55*103 1.37*103

The results of the sensitivity analysis for the individual one litre wine packages indicate that the transportation distance substitutions can increase net life cycle endpoint impacts by up to 9% and reduce them by up to 8% (Table 10). The effects of the distance substitutions are far more pronounced for the CSU and LSU glass containers than for the other lighter weight containers.

Table 10 Percent change from reference net endpoint impact values for individual wine containers

Endpoint Percent change from reference net endpoint impact values, by type of wine package impact AC CSU LSU PET RFG Container transportation distance to packager based on subtraction of weighted standard deviations Ecosystem -1.1% -5.1% -5.0% -1.4% -2.6% diversity Human health -2.7% -6.2% -6.2% -1.9% -3.0% Resource -3.0% -8.2% -8.1% -1.5% -3.6% availability Container transportation distance to packager based on addition of weighted standard deviations Ecosystem 1.2% 5.8% 5.7% 1.6% 3.0% diversity Human health 3.2% 7.5% 7.4% 2.2% 3.6% Resource 3.5% 9.4% 9.2% 1.7% 4.0% availability

For the 2008 reference scenario at the municipal scale, the sensitivity analysis results displayed in Table 11 demonstrate that the proportional effects of the substituted transportation distances on the net life cycle impacts tend to be even larger than those of the individual package LCA for the one litre CSU glass container (the heaviest container in the individual package LCA comparison, responsible for greater transportation emissions than the alternative packages) (Table 10). This finding would seem to be 107

counterintuitive because the municipal scale 2008 reference scenario comprises many types of containers, including those in which the transportation distance substitution effects are considerably smaller, including ACs and PET containers. However, the 2008 reference scenario not only includes containers of different types, but containers of different sizes. Smaller containers require more material per unit of wine/spirit volume held than do larger containers, and both of the municipal scale scenarios have a greater volume of wines and spirits packaged in containers smaller than one litre than in those larger than one litre (see Table 4 and Appendix 3). Therefore, it is not surprising that, for the 2008 reference scenario, the net impact changes due to the transportation distance substitutions tend to be proportionally higher than in the one litre container scenarios. In comparison, the alternative packaging scenario shows smaller proportional changes than the 2008 reference because it has a much larger market penetration of lighter weight packages which are responsible for lower transportation emissions than conventional single use glass bottles.

Table 11 Percent change from reference net endpoint impact values for 2008 reference and alternative packaging scenarios

Endpoint impact 2008 reference scenario Alternative packaging scenario Container transportation distance to packager based on subtraction of weighted standard deviations Ecosystem diversity -5.1% -4.7% Human health -6.4% -5.9% Resource availability -8.1% -7.3% Container transportation distance to packager based on addition of weighted standard deviations Ecosystem diversity 5.9% 5.4% Human health 7.7% 7.1% Resource availability 9.3% 8.3%

4.7 Critical review

Critical reviews are undertaken to evaluate an LCA’s scientific and technical validity, as well as to review the completeness and consistency of the LCA inputs and results (ISO 2000). Part of such reviews is an examination of the sources of uncertainty in terms of the precision and accuracy of the input data and unit processes, as well as the LCA method applied. LCA models incorporate LCA profile data (e.g., the tonnage of glass produced), LCA unit process data (e.g. energy inputs for glass production), emission and resource use data, as well as LCIA characterization and modelling data. 108

Where possible, primary data such as the questionnaire responses by wine and spirit suppliers were used as inputs to the LCA unit processes. The low response rates to the questionnaires and their incomplete geographic coverage introduced considerable uncertainty to the mean transportation distance estimates. This uncertainty was limited to only a portion of the packaging transportation component of the life cycle (package manufacturer to packager). The sensitivity analysis in Section 4.6 explored the consequences of this uncertainty by evaluating the life cycle results as a function of various transportation distances (i.e., adding and subtracting the weighted standard deviations). I am unaware of any recent published source of this type of data that could be considered more representative of wine and spirit packaging transportation. In comparison with the questionnaire responses, far less uncertainty was associated with the mass measurements of the various types of containers, secondary packaging and spirit brands. In addition to using direct measurements, uncertainty was minimized by using the unpublished LCBO 2008 volume sales data. The sales data reduced the uncertainty of the average masses of each type of container, closure and capsule by permitting the cross-referencing of the mass measurements with the actual sales volumes by brand, container volume, as well as type of container and closure used (see Appendix 3). The standard deviation of the container masses ranged from less than 1% to 14% of the average. The masses of the containers were measured using a digital kitchen scale with a measurement readability of 1 g. The accuracy of this scale was confirmed to be within 1 g by weighing the secondary packaging using both the kitchen scale and a Sartorius Research balance which had a measurement readability of 10-5 g and a measurement standard deviation less than or equal to ± 0.05 mg. The measurements of the masses of the secondary packaging (to 10-1 g) and spirits (to 10-3 g) were taken using the Sartorius Research balance. The 2008 LCBO volume sales data were also used to ensure that the municipal scale 2008 reference scenario took into account the best available data for overall consumption of wine/spirit packaging in terms of the material composition of the primary and secondary packaging. Some uncertainty was introduced by assuming that wine and spirit consumption by Toronto residents (in terms of the relative popularity of brands and 109

container types) did not differ significantly from the consumption patterns of the remaining 80% of Ontarians. Some uncertainty was also associated with the selection and design of the unit processes included in the LCA model to represent container and secondary packaging production, transportation and waste management processes. In most cases, the unit processes for this LCA were derived from the US-EI database, and were based on the technology used in Switzerland, while using the United States electricity mix. I modified these unit processes in order to improve their representation of the system studied relative to technology employed and data age (e.g., bottle washing, glass production, recycling), as well as electricity mix used. The most significant unit processes contributing to the endpoint impacts were related to glass production. The unit processes for glass included several modifications of the EcoInvent unit processes that were proposed in a comprehensive LCA of Quebec (Canada) beer containers by Magaud et al. (2010). These modifications permitted the glass unit processes to specify the recycled content and the means of glass production. For example, the unit process for glass now depicts glass production using an electric boost furnace, which is used in the majority of glass production facilities in the United States) (Magaud et al. 2010). Although the amount of energy used per tonne of container glass produced is approximately 6.4 GJ using the LCA unit process, the energy use of container glass production can vary from 4 to 10 GJ/tonne (Beerkens and van Limpt 2001). Nevertheless, as the glass wine and spirit containers examined in this LCA were produced in numerous facilities across the globe, it seems unlikely that the mean energy requirements for the production of the glass containers would be close to either extreme (i.e., 4 or 10 GJ/tonne). Capital/infrastructure accounting was not always incorporated into the unit processes due to a lack of available data, thereby generating some system boundary inconsistencies. Omissions included facilities and equipment for PET recycling, bottle washing and filling, and aseptic carton manufacturing. However, since these three unit processes, with the possible exception of PET recycling (see Figures 14 to 16), were responsible for very small portions of the net endpoint level life cycle impacts, it seems unlikely that the capital/infrastructure omissions would generate significant effects on the 110

LCA results. Similarly, the omission of the material and energy inputs for the container rinsing and filling processes did not significantly impact the results since the bottle washing process contributed no more than 0.4% of any of the life cycle endpoint level impacts for the municipal scale scenarios. Bottle washing contributed only 2% of the impacts for the RFG individual package scenario. Uncertainty in the final results was also due to uncertainties introduced during the impact assessment stage of LCA, in which life cycle emissions were classified and weighted to produce midpoint and endpoint level impact indicators. Moreover, uncertainty increases significantly when interpreting results using endpoint rather than midpoint impacts (Goedkoop et al. 2009) because endpoint level indicators require an additional weighted valuation of impacts. ReCiPe groups the uncertainties associated with some of its impact conversion and aggregation measures into “perspectives” (egalitarian, hierarchist, individualist) (Goedkoop et al. 2009). In order to provide an indication of the uncertainty demarcated by these three perspectives, LCIA results for individual packaging systems were compared using ReCiPe’s three perspectives. Di Maria and Fantozzi (2004) conducted a similar type of comparison of results for a waste management LCA using the three perspectives of Eco-lndicator 99. For individual wine containers, the uncertainty analysis addressed the percentage differences in the net endpoint level impacts between each alternative container system and the CSU glass container. This analysis indicated that uncertainty is very low for the damage to resource availability indicator, with a relative difference of less than 1% between the impacts using the hierarchist perspective in comparison with the other two perspectives. In contrast, the relative difference for the damage to human health indicator was up to 37%, while that for damage to ecosystem diversity reached 28%. The choice of perspective had no effect on the ranking of each packaging system, except for the ranking of RFG wine containers and aseptic cartons using the damage to human health indicator. However, this exception was more likely because the potential net impacts of these two container systems were of similar magnitude (see Table 8). In addition to the uncertainty conferred by LCIA methodological assumptions, indicators of damage inherently include assumptions pertaining to the geographical and ecological character of the locations of the impacts. The influence of spatial variation on 111

the uncertainty of the results was not evaluated in this LCA due to (1) the relative underdevelopment of methods to regionalize impacts within LCIA (Mutel and Hellweg 2009); and (2) the environmental impacts depicted in the LCA scenarios were broadly distributed geographically as a consequence of the substantial number of locations of container and secondary packaging manufacture. Overall, the uncertainty associated with the key data inputs such as container mass, container type, and transport distances, was relatively low. The LCA inputs were fairly representative of the modeled systems in terms of (1) the cross-referencing of the container and secondary packaging mass measurements with 2008 volume sales data from the LCBO which depicted the brand types, container volumes, and container and closure types; and (2) the accuracy and precision of the measurements of the masses of the containers, secondary packaging and spirits. Additional uncertainty was associated with the LCA unit processes selected and designed to represent the actual processes subject to modelling. Although this uncertainty is substantial, it would not be sufficient to alter the ranking of the containers in terms of impacts, with the exception of the rankings of ACs and RFG containers, which have very similar endpoint impact levels. The relatively high uncertainty of container transportation distances from the manufacturer to the packager was addressed in the sensitivity analysis, the results of which did not affect the original endpoint level results by more than 10%. The impact assessment stage also introduces considerable uncertainty to the results, especially for human health and certain ecological impacts. However, the endpoint level impacts of each individual package scenario relative to one another did not differ significantly using the three different ReCiPe perspectives.

4.8 Discussion

The results of this LCA indicate that the net environmental burdens from each package life cycle broadly reflect the relative masses of the containers, with the exception of refillable glass bottles. They show the refillable glass bottle and aseptic carton as having the lowest potential net environmental impacts, responsible for up to an 87% reduction in endpoint level impacts relative to the CSU glass bottle. 112

The environmental performance of the packaging system is sensitive to the distance travelled to market, with reductions in the average distances improving the overall performance of the heavier packages (CSU and LSU glass bottles) to a greater extent than they would the lighter ones. Unless there would happen to be substantial increases in bottle breakage from using particular designs of LSU glass bottles, LSU glass bottle substitutions for CSU glass containers would only reap environmental and human health benefits. For refillables, a focus on reducing secondary packaging impacts (for wine containers), and/or increasing the average number of bottle reuses, could result in a refillable container system with much lower impacts than any of the packaging alternatives. At the municipal scale, the alternative wine/spirit packaging scenario can reduce packaging waste by approximately half, and potential net impacts by somewhat less than half (40%). As was discussed in Section 4.2.2.2, not all of the package types are fully substitutable. For example, 750 ml aseptic cartons do not yet exist, and aseptic cartons should not be used for sparkling wines. Therefore, it would not be appropriate to switch entirely to one particular form of wine/spirit package. Should the package substitutions associated with the alternative wine/sprit packaging scenario be applied across the province of Ontario, the LCBO goal of reducing waste by 10 million kg annually would be exceeded by approximately five times. Although LCAs of individual packages are used to estimate the potential net environmental gains from package substitutions, they can be misleading if one’s interpretation and extrapolation of the results lacks an appropriate context. For example, it is not realistic to assume that human health impacts can be reduced by a factor of six by substituting one litre aseptic cartons for all CSU glass containers. Undertaking the packaging LCA at a municipal scale provides a much more realistic appraisal of the potential environmental gains from package substitutions because, at this scale, one can address levels of product consumption, including variations in product types consumed (i.e., the sizes and types of wine/spirit packages).

4.9 Conclusion

113

The substitution of lightweight and refillable packages for conventional glass containers has considerable potential to reduce the environmental impacts of wine and spirit packaging. The LCAs of the individual packaging systems and the packaging used by the residents of the City of Toronto, Canada illustrate the differing means by which one may evaluate product systems. The former type of LCA (individual package scale) does not address the actual levels of product consumption, and the potential environmental gains from multiple types of functionally-equivalent package substitutions that would take place in one specified region. The latter type of LCA (municipal scale) does not supply results from which one can derive the environmental performance of individual packaging systems. Together, the results from these LCAs can provide useful input for commercial procurement decisions, policy design, product design and stewardship, and waste management.

4.10 References

Association of Canadian Distillers. 2008. Annual statistics.

Bare J, Norris G, Pennington D, McKone T. 2003. TRACI: The Tool for the Reduction and Assessment of Chemical and Other Environmental Impacts. J Ind Ecol 6(3-4): 49- 78.

Bavcar D, Komerl T. 2002. Determination of Alcohol and Dry Extract in Wines. Food Technol. Biotechnol 40(4):321–329.

Beerkens R, van Limpt J. 2001. Chapter 7: Energy efficient benchmarking of glass furnaces. 62nd Conference on Glass Problems: Ceramic Engineering and Science Proceedings. Volume 23. Issue 1. University of Illinois. DOI: 10.1002/9780470294727.ch7.

Bengoa X, Maia De Souza D, Samson R. 2009. Evaluating Tradeoffs between Material Supply, Lightweighting and Recyclability using Life Cycle Assessment: A Case Study on Wine Packaging. Presentation. LCA IX Boston. Available at: http://www.lcacenter.org/LCA9/presentations/212.pdf. Accessed on 2011 06 08.

Canadian Vintner’s Association. 2008. Annual statistics.

City of Toronto. 2010a. Personal communication. 2010 06 30.

114

City of Toronto. 2010b. City of Toronto: Solid Waste Management - Green Bin Program. Frequently asked questions. Available at: http://www.toronto.ca/greenbin/faq.htm. Accessed on 2010 08 21.

City of Toronto. 2008. 2008 breakdown by material of residential waste diversion. Available at: http://www.toronto.ca/garbage/pdf/2008-chart.pdf. Accessed on 2010 08 21.

City of Toronto. 2007. Proposed Initiatives and Financing Model to Get to 70% Solid Waste Diversion by 2010. Prepared for the City of Toronto Executive Committee by City of Toronto Solid Waste Management Services. Reference No. p:/2007/swms/may/011EC.doc. Available at: http://www.toronto.ca/legdocs/mmis/2007/ex/bgrd/backgroundfile-3799.pdf. Accessed on 2010 09 15.

City of Toronto. 2006. Backgrounder: Release of 2006 Census results: Population and Dwelling Counts. Available at: http://www.toronto.ca/demographics/pdf/2006_population_and_dwelling_count_backgro under.pdf. Accessed on 2010 09 15.

CM Consulting. 2010. Who pays what: An analysis of beverage container recovery and costs in Canada. Available at: http://www.bcmb.ab.ca/pdf/Who_Pays_What_2010.pdf. Accessed on: 2011 03 03.

Constar. 2010. MonOxbar®. Available at http://www.constar.net/tech-barrier- monoxbar.php. Accessed on 2010 07 01.

Cox, J; Giorgi, S; Sharp, V; Strange, K; Wilson, D; Blakey, N. 2010. Household waste prevention — a review of evidence. Waste Manag Res 28: 193-219.

Ecoinvent Centre. 2004. Ecoinvent data v1.1. Final reports Ecoinvent 2000 (1–15). Swiss Centre for Life Cycle Inventories, Duebendorf (CD-ROM)

Ekvall T, Assefa G, Bjorklund A, Eriksson O, Finnveden G. 2007. What life-cycle assessment does and does not do in assessments of waste management. Waste Manage 27: 989-996.

Fisher, C. 2010. 2010 Capsule Report: Small Wineries Maintain Course with Tin Capsules Despite Skyrocketing Costs. Wine Business Monthly. September 15, 2010. Available at: http://www.winebusiness.com/wbm/?go=getArticle&dataId=80421. Accessed on 2011 05 27.

Franklin Associates. 2010. Final Report. Life cycle inventory of 100% postconsumer HDPE and PET recycled resin from postconsumer containers and packaging. Available at: http://www.americanchemistry.com/s_plastics/sec_pfpg.asp?CID= 1439&DID=10907. Accessed on 2010 10 10. 115

Franklin Associates. 2006. Life Cycle Inventory of Container Systems for Wine. Final Report. Prepared for Tetra Pak Inc. Available at: http://www.tetrapak.ca/pics/winepack- report-EN.pdf. Accessed on 2010 04 06.

Frischknecht R, Althaus H-J, Bauer C, Doka G, Heck T, Jungbluth N, Kellenberger D, Nemecek T. 2007a. The Environmental Relevance of Capital Goods in Life Cycle Assessments of Products and Services. Int J LCA, DOI: http://dx.doi.org/10.1065/lca2007.02.308.

Frischknecht R, Jungbluth N, Althaus H.-J, Doka G, Heck T, Hellweg S, Hischier R, Nemecek T, Rebitzer G, Spielmann M, Wernet G. 2007b. Overview and Methodology. Ecoinvent report No. 1. Dübendorf, Switzerland: Swiss Centre for Life Cycle Inventories. 68 pp. Available at http://www.ecoinvent.org/fileadmin/documents/en/ 01_OverviewAndMethodology.pdf. Accessed on 2008 12 07.

Frischknecht R, Jungbluth N, Althaus H.-J, Bauer C, Doka G, Dones R, Hischier R, Hellweg S, Humbert S, Köllner T, Loerincik Y, Margni M, Nemecek T. 2007c. Implementation of Life Cycle Impact Assessment Methods. ecoinvent report No. 3, v2.0. Dübendorf, Switzerland: Swiss Centre for Life Cycle Inventories. Available at http://www.ecoinvent.org/fileadmin/documents/en/03_LCIA-Implementation.pdf. Accessed on 2008 12 07

Goedkoop M, Heijungs R, Huijbregts M, De Schryver A; Struijs J, van Zelm R. 2009. ReCiPe 2008: A life cycle impact assessment method which comprises harmonised category indicators at the midpoint and the endpoint level. First edition. Report I: Characterisation. Rumte en Milieu. Ministerie van Volkshuisvesting. Ruimtelijke Ordening en Milieubeheer. Available at: http://s3.amazonaws.com/jef.mindtouch.com/10059895/11/0?AWSAccessKeyId=1TDEJ CXAPFCDHW56MSG2&Signature=/M2Jz8%2b3G0vMifS6QJM3bieLwRc%3d&Expir es=1282315930. Accessed on 2010 08 20.

Google. 2010. Google Maps. Available at: http://maps.google.ca/maps. Accessed on 2010 02 18.

Hung M-L, Ma H-w. 2009. Quantifying system uncertainty of life cycle assessment based on Monte Carlo simulation. Int J LCA 14(1): 19-27.

Jolliet O, Margni M, Charles R, Humbert S, Payer J, Rebitzer G, Rosenbaum R. 2003. IMPACT 2002+: A new life cycle impact assessment methodology. Int J LCA 8(6): 324-330.

Koel, J. 2008. Personal communication with J. Koel. Tetra Pak Canada. Richmond Hill, Ontario, Canada, 2008 12 15.

116

ICF Consulting. 2005. Determination of the Impact of Waste Management Activities on Greenhouse Gas Emissions: 2005 Update. Toronto, Ontario: ICF Consulting. Submitted to: Environment Canada and Natural Resources Canada. Contract No. K2216-04-0006. Available at: http://mmsd1.mms.nrcan.gc.ca/recycle/ICF%20final%20report.pdf. Accessed on 2010 08 21.

Independent Electricity System Operator (IESO). 2009. IESO 2008 Electricity Figures Show Record Levels of Hydroelectric Power. News Release. Available at: http://www.ieso.ca/imoweb/media/md_newsitem.asp?newsID=4458. Accessed on 2010 06 27.

Institut des sciences de la vigne et du vin (ISVV). 2010. Influence of packaging on wine preservation. 12 month study, 2009-2010. PowerPoint Presentation. Available at: http://www.slideshare.net/LisaTwidell/does-the-packaging-of-the-future-match-the- future-needs-of-wine. Accessed on 2010 06 06.

International Energy Agency. 2010. IEA Energy Statistics – For Electricity/Heat (2008). Available at: http://www.iea.org/stats/prodresult.asp?PRODUCT=Electricity/Heat. Accessed on 2010 12 27.

International Organization of Standardization. 2006 Environmental management - Life cycle assessment - Requirements and guidelines. ISO 14044: 2006.

Laursen, S. E., Hansen, J., Bagh, J., Jensen O. and Werther, I., Environmental Assessment of Textiles. Life Cycle Screening of Textiles containing Cotton, Wool, Viscose, Polyester or Acrylic Fibres, Miljøstyrelsen Miljøprojekt nr. 369, DTI Clothing and Textile, dk-TEKNIK, Danmark, 1997.

Leighton, C. 2010. Refillable wine bottles. Solid Waste & Recycling. October/November 2010 issue: 8-16. Available at: http://www.solidwastemag.com/issues/de.aspx?id=5474. Accessed on 2011 02 05.

Liquor Control Board of Ontario. 2008. 2008 Volume sales data for wine and spirits. Unpublished.

Liquor Control Board of Ontario. 2008. Responsible Retailing Paying Dividends: LCBO Annual Report 2007–08. Toronto, Ontario, LCBO. 60 pp.

Liquor Control Board of Ontario. 2006. LCBO Annual Report 2005-2006. Toronto, Ontario: LCBO Corporate Communications. 46 pp.

Liquor Control Board of Ontario. n.d. LCBO – Enviro Report: Alternative packaging. Available at: http://www.lcbo.com/socialresponsibility/enviroreport/wmv/alt_packaging.wmv. Accessed on: 2010 07 13. 117

Magaud V, Van Durme G, Bengoa X. 2010. Rapport final. Analyse du cycle de vie de contenants de bière au Québec. Centre Interuniversitaire de recherché sur le cycle de vie des produits, procédés, et services. Available at: http://www.recyc- quebec.gouv.qc.ca/Upload/publications/consigne/2010/Analyse-cycle-biere-rap.pdf. Accessed on 2011 03 02.

McDougall F, White P, Franke M, Hindle P. 2001. Integrated Solid Waste Management: a Life Cycle Inventory. Second Edition. Oxford, United Kingdom: Blackwell Publishing.

Moody B. 1977. Packaging in Glass. London, England: Hutchison Benham. 383 pp.

Mourad A, Garcia E, Vilela G, von Zuben F. 2008. Environmental Effects from a Recycling Rate Increase of Cardboard of Aseptic Packaging System for Milk Using Life Cycle Approach. Int J LCA 13(2): 140–146.

Mutel C, Hellweg S. 2009. Regionalized Life Cycle Assessment: Computational Methodology and Application to Inventory Databases. Environ Sci Technol 43:5797- 5803.

National Association for PET Container Resources (NAPCOR). 2008. 2008 Report on Postconsumer PET Container Recycling Activity. Available at: http://www.napcor.com/pdf/2008_Report.pdf . Accessed on 2011 03 04.

Owens-Illinois. 2009. Consumers Prefer Wine in Glass to Alternative Packaging. http://www.o-i.com/uploadedFiles/Web_Site/O- I/NorthAmerica/MARKET_CATEGORIES/Wine/O- I%20Wine%20Global%20Consumer%20Research%20Release.pdf. Accessed on 2009 11 08.

PricewaterhouseCoopers (PwC)/ECOBILAN. 2008. Evaluation of the environmental impacts of Cork Stoppers versus Aluminium and Plastic Closures: Analysis of the life cycle of Cork, Aluminium and Plastic Wine Closures. 126 pp. http://www.corkfacts.com/pdffiles/Amorim_LCA_Final_Report.pdf. Accessed on 2009 11 08.

Rigamonti L, Grosso M, Caterina M. 2009. Influence of assumptions about selection and recycling efficiencies on the LCA of integrated waste management systems. Int J LCA 14: 411–419.

Rockwell Automation Inc. 2010. Application solutions: Casella Wines deploys MES solution tracking downtime. Available at http://www.rockwellautomation.com.au /applications/gs/ap/gsau.nsf/pages/Casella_Wines_bottling_line. Accessed on 2010 07 04.

118

Salhofer S, Obersteiner G, Schneider F, Lebersorger S. 2008. Potentials for the prevention of municipal solid waste. Waste Manage 28: 245-259.

Saphire D. 1994. Case Reopened: Reassessing Refillable Bottles. New York, New York: Uniform Inc. 351 pp.

Spielmann, M., Bauer, C., Dones, R., Tuchschmid, M. 2007. Transport Services: ecoinvent report No. 14. Swiss Centre for Life Cycle Inventories, Dübendorf.

Statistics Canada. 2008. CHASS Trade Analyser: Canadian Imports Data. 21 Dec 2008. HS10 Codes Used for Imported Bottled Wines: 2204101000, 2204109000, 2204211010, 2204211091, 2204211092, 2204211099, 2204212100, 2204212200, 2204212310, 2204212390, 2204212410, 2204212420, 2204212490, 2204212510, 2204212520, 2204212590, 2204212610, 2204212620, 2204212630, 2204212690, 2204212710, 2204212720, 2204212790, 2204212800, 2204213110, 2204213120, 2204213130, 2204213190, 2204213210, 2204213220, 2204213230, 2204213290, 2205101010, 2205101020, 2205101030, 2205101090, 2205102000, 2205103000, 2206002130, 2206002200, 2206003100, 2206003900, 2206004100, 2206004900, 2206005011, 2206005019, 2206005090, 2206006100, 2206006200, 2206006300, 2206006410, 2206006490, 2206006500, 2206006600, 2206006700, 2206006800, 2206007100, 2206007200.

HS10 Codes Use for Imported Wine in Bulk (containers larger than 2L) 2204291010, 2204291020, 2204291090, 2204292100, 2204292200, 2204292300, 2204292400, 2204292500, 2204292600, 2204292700, 2204292800, 2204293100, 2204293200, 2204301000, 2204309000, 2205901000, 2205902000, 2205903000

HS10 Codes Used for Imported Bottled Spirits 2208200090, 2208300019, 2208300029, 2208300030, 2208300040, 2208300099, 2208401090, 2208409000, 2208500090, 2208600000, 2208700000, 2208901000, 2208903000, 2208909200.

HS10 Codes Used for Imported Spirits in Bulk (containers larger than 2L) 2208200010, 2208300011, 2208300021, 2208300091, 2208401010, 2208500010, 2208902100, 2208902900, 2208909900 -Assumption that undenatured ethyl alcohol is shipped in bulk

Statistics Canada. 2006. Population and dwelling counts, for Canada, provinces and territories, and census subdivisions (municipalities), 2006 and 2001 censuses - 100% data. Census of Population. Retrieved on 2007 03 13.

Stewardship Ontario. 2006a. Multi-Family Waste Audit Program. Toronto (Downtown): Summer 2005, Fall 2005, Winter 2006, Spring 2006 (assumed 50% of multiple unit dwellings). Toronto (Scarborough) Multiple Unit Dwellings: Winter 2006, Spring 2006, Summer 2006, Fall 2005 (assumed 50% of multiple unit dwellings). Available at: 119

http://www.stewardshipontario.ca/eefund/projects/audits/waste_audit_mf.htm. Accessed on 16 March 2009.

Stewardship Ontario. 2006b. Single Family Waste Audit Program. Toronto Single Family Dwellings: Winter 2005, Spring 2005, Summer 2005, Fall 2005. Available at: http://www.stewardshipontario.ca/eefund/projects/audits/waste_audit_sf.htm. Accessed on 16 March 2009.

Stewardship Ontario. 2005. Ontario Blue Box Markets Overview. 2002/2003 Blue Box Materials Generated and Municipally Marketed. Mass Balance Report. Available at: http://www.wdo.ca/files/domain4116/OBFR.pdf. Accessed on 2010 02 27.

Tetra Pak. 2003. L'engagement de Tetra Pak pour le développement durable : France- Belgique. Edition 2003. Available at: http://www.corporateregister.com/a10723/ Tetra03-sus-fr.pdf. Accessed on 2010 08 26.

Tetra Pak. 2005. The Recycling Brochure for Europe. Available at: http://www.tetrapak.com/docs/recycling_booklet.pdf. Accessed on 2008 01 20.

The Beer Store (TBS). 2009. Design for the environment: Responsible Stewardship 2008-2009. Available at: http://www.wdo.ca/files/domain4116/ Responsible%20Stewardship%2008-09_FINAL.pdf. Accessed on 2011 06 08.

Toffoletto L, Bulle C, Godin J, Reid C, Deschênes L. 2007. LUCAS – A New LCIA Method Used for a CAnadian-Specific Context. Int J LCA 12(2):93-102.

Toronto City Council. Works Committee / Toronto City Clerk. 2002. Blue Box Residue and Recycling of Coloured Glass. Report No. 6. Clause No. 2. Available at http://www.toronto.ca/legdocs/2002/agendas/council/cc020521/wks6rpt/cl002.pdf. Accessed on 2008 02 20.

Valiante U. 2007. A Look at Tetra Pak's New Life-Cycle Inventory. Solid Waste & Recycling. February/March 2007.

Waste Diversion Ontario. 2008. Municipal Datacall: Tonnage. Available at: http://www.wdo.ca/content/?path=page82+item35931. Accessed on 2010 05 29.

Waste & Resources Action Programme (WRAP). 2008. The carbon impact of bottling Australian wine in the UK: PET and glass bottles. Final Report. Project code: RTL063- 001. Available at: http://www.wrap.org.uk/downloads/Carbon_Impact_of_Bottling_Australian_Wine_in_th e_UK-_PET_and_Glass_Bottles.79406545.pdf. Accessed on 2008 05 06.

World-Ports Distances Calculator. 2010. Available at http://www.distances.com/. Accessed on 2010 08 07.

120

PAPER 4

Waste prevention and life cycle assessment of residential waste management in Toronto, Canada

121

5.1 Introduction

Life cycle assessments (LCAs) of municipal solid waste (MSW) management systems generally ascribe no environmental burden to processes leading to the generation of waste in order to simplify their system boundaries and thus streamline the assessment. This “zero burden assumption” prohibits the comparison of MSW management scenarios which incorporate waste prevention, including product reuse, on a functionally equivalent basis. The widespread public policy emphasis on increasing waste prevention and diversion (e.g., European Commission 2005) has created a need for LCAs that account for the burdens and avoided burdens associated with all forms of MSW management, including not only waste treatments, but also waste prevention activities (WPAs). Although rarely addressed in published LCAs of MSW, authors such as Coleman (2003), Ekvall et al. (2007) and Cleary (2010 / Chapter 3), have examined the methodological issues associated with incorporating waste prevention into this type of LCA. There has also been a notable increase in academic publications on waste prevention in the waste management literature (e.g., Cox et al. 2010; Salhofer et al. 2008). Through the modification of the “zero burden approach,” one may compare within one system boundary the various components of the waste management hierarchy, including WPAs and waste treatments, so that they are regarded as functional equivalents in managing MSW. Cleary’s (2010 / Chapter 3) Waste Management and Prevention (WasteMAP) LCA is such a conceptual model. Since an LCA of a residential waste management system using WasteMAP has not yet been undertaken, this paper is an attempt to do so. The City of Toronto, Canada, with a population of approximately 2.5 million, is an example of a municipality which includes some WPAs in its calculation of the quantity of waste diverted from landfills. Indeed, waste diversion is considered to be a prime indicator of the success of Toronto’s MSW management system, and the City has committed itself to diverting 70% of its residential waste from landfill over the next few years (City of Toronto 2007). In order to help achieve this goal, Toronto has increased the variety of materials that are recycled, and has introduced a source-separated organic waste treatment program. Educational programs and financial incentives have also been 122

employed to encourage waste prevention activities, such as the “City of Toronto Municipal Code Chapter 604, Packaging,” which requires retailers to charge five cents or more for each plastic shopping bag supplied to customers (City of Toronto 2011). Although many types of WPAs can contribute to residential waste management, the only WPAs taken into account in the City’s waste diversion calculation are backyard composting and grasscycling, both forms of on-property residential waste treatment. The waste prevention goals of this municipality make the City of Toronto ideal for an LCA of residential waste management using the system boundary and functional unit assumptions specified in the WasteMAP LCA model. The results of a study by Salhofer et al. (2008) indicated that, relative to total quantities of residential waste generated in Vienna, Austria, the apparent potential of waste prevention measures to reduce waste generation is generally small. Their measures, addressing waste from advertising mail, beverage packaging, diapers, food waste, and ‘big events,’ were capable of reducing the size of each waste stream by approximately 10%, equal to 1-3% of all municipal waste (Salhofer et al. 2008). However, it was not within the scope of their study to evaluate the net environmental benefits of these WPAs through an LCA. The following case study of the WasteMAP LCA model is a response to the relative dearth of information about the environmental impacts of residential waste management systems that incorporate waste prevention.

5.2 Research objectives

The research objectives of this paper are (1) to evaluate the net environmental burdens of the residential solid waste management system of the City of Toronto, Canada; (2) to apply the WasteMAP LCA model (Cleary 2010 / Chapter 3) to estimate the environmental performance of an alternative, functionally equivalent scenario in which a lower quantity of waste is generated due to the incorporation of waste prevention activities; and (3) to use the LCA results to assess the potential of a broad set of waste prevention activities to reduce the net life cycle impacts of this urban waste management system. Moreover, the functionally equivalent LCA scenarios (see Section 5.3.1) are 123 compared with one another to evaluate the environmental significance of various WPAs relative to the other components of MSW management life cycle.

5.3 Methodology

This LCA focuses on residential waste managed in the City of Toronto, based on 2008 data. The scenarios examined not only account for those processes associated with the collection, sorting and treatment of residential waste inputs, but also reflect the prevention of waste materials through various types of WPAs. These WPAs include (1) the reduced generation of unaddressed advertising mail; (2) the reuse of disposable carry- out plastic shopping bags; (3) the substitution of news articles available online for those printed on newsprint; (4) the substitution of reusable glass wine and spirit bottles for single use ones; (5) the substitution of lightweight glass bottles, plastic bottles and aseptic cartons for conventional wine and spirit containers; and (6) grasscycling. Each of these WPAs exemplifies one of the main types of WPAs listed in Cleary (2010 / Chapter 3) and depicted in Table 3. Since waste from durable goods is outside of the system boundary of the LCA (see Section 5.3.3), an example of WPA-6, lengthening the useful lifespan of a durable good, has been omitted.

5.3.1 Residential waste management scenarios

Two residential waste management scenarios are evaluated and compared using the WasteMAP LCA model. The first is the 2008 reference scenario for the City of Toronto’s residential waste management system, in which the amount and composition of residential waste managed, is identical to the amount of waste generated and collected. In contrast, the alternative to the 2008 reference, designated the waste prevention scenario, incorporates into the waste management system the net effect of implementing six types of WPAs, which are considered waste management processes that are undertaken to reduce net waste generation relative to the 2008 reference. The WPAs in the waste prevention scenario, and the corresponding quantities of waste prevention, are listed and described in Table 12. 124

Table 12 The waste prevention activities addressed in the waste prevention scenario, the properties of each type of WPA and the quantities of waste prevention

Waste prevention Type of WPA (based upon Description of WPA and quantity of waste prevention1 activities (WPAs) Cleary 2010 / Chapter 3)

AdMail WPA WPA-1 Reduction in Reduced generation of unaddressed advertising mail from Reduced generation of material consumption private carriers (excluding advertising inserts in subscription unaddressed advertising without product service newspapers), equal to the difference between the 2% already mail substitution opting out of admail, and the 67% of Canadians who find admail to be of no interest

[10,007 tonnes of prevented newsprint waste] Disposable Bag WPA WPA-2 Reuse of a Reuse of 20% of disposable carry-out plastic shopping bags Reuse of disposable carry- disposable good out plastic shopping bags [1,216 tonnes of prevented HDPE waste] Newspaper WPA WPA-3 Substitution of a Substitution of articles available online for those printed on Substitution of articles service, provided by a newsprint, resulting in the prevention of 10% of newsprint from available online for those capital good, for a Toronto daily subscription newspapers printed on newsprint disposable good [4,528 tonnes of prevented newsprint waste, and 10 tonnes of additional household computer equipment] Refillable Bottle WPA WPA-4 Substitution of a Refillable glass (RFG) bottles are substituted for all 750 ml, Substitution of refillable reusable good for a 1000 ml, and 1500 ml containers for Canadian and bulk glass bottles (RFGs) for disposable one imported wines in 2008, as well as 750 and 1140 ml containers conventional single use for Canadian and bulk imported spirits (CSU) wine and spirit bottles Lightweight Container WPA-5 Lightweighting of a (1) Aseptic cartons (ACs) are used to package 10% of the 2008 WPA good volume sales of imported packaged wines in 1500 ml Substitution of lightweight containers, and 50% of the 2008 volume sales of imported single use (LSU) glass packaged wines in 1000 ml containers bottles, PET plastic bottles (2) PET bottles replace 10% of the 2008 volume sales of and aseptic cartons (ACs) imported packaged spirits in 750 ml and 1140 ml containers, for conventional and 50% of the 2008 volume sales of imported packaged spirits containers2 in 200 ml, 375 ml, 1750 ml and the remaining container sizes (3) The remaining containers, excluding RFG bottles, are packaged in lightweight single use (LSU) bottles (assumption of 20% lower mass than CSU containers, for all container sizes)

[9,248 net tonnes of prevented packaging wine and spirit packaging waste due to refillable bottles and lightweight containers] N/A WPA-6 Lengthening the N/A useful lifespan of a durable good Grasscycling WPA WPA-7 On-property An additional 5% of the mass of yard waste is prevented due to Grasscycling residential waste treatment grasscycling

[4,028 tonnes of prevented yard waste] 1 Although for calculation purposes in the model the waste prevention estimates are rounded to the tonne, the specific numbers of digits associated with all of these estimates are not intended to indicate a particular level of certainty (uncertainty is addressed in Section 5.8). 2 PET plastic wine containers with volumes of 375 ml, 1.5 L, 2 L and 3 L PET do not yet exist, or did not exist in 2008. Since only wine containers with volumes of 750 ml or larger were a significant portion of the market in 2008 (>1%), only 1 litre and 1.5 litre aseptic cartons are incorporated into the waste prevention scenario. LSU glass bottles are assumed to have a 20% lower mass than CSU containers, for all container sizes. Please see Cleary (unpublished / Chapter 4) for more details on wine and spirit packaging data.

125

The unaddressed advertising mail (admail) WPA is selected to exemplify WPA-1 because the aim of this measure is to prevent the production of unwanted products that generate waste. This type of WPA does not affect the functional equivalence of the LCA scenario comparison with the 2008 reference because only the unwanted admail is prevented. There is no reduction in product service consumption. In order to generate a plausible estimate of the potential for admail prevention, it is necessary to take into account recent levels of admail refusal and the percentage of the population that does not wish to receive this material. Some voluntary initiatives have been introduced in North America (e.g., Red Dot Campaign n.d.) to foster the reduction of admail, including admail opt-out provisions by Canada Post (n.d.). However, these initiatives have not been successful in reducing admail to an extent comparable to the surveyed percentage of the Canadian population who find this advertising of no interest to them, which has been estimated at 67% (FDSA 2007; based on results from a telephone survey commissioned by the Canadian Marketing Association in 2005). According to Canada Post, only two percent of Canadians have opted out of receiving admail (Canada Post 2007). However, this institution claims to be responsible for only 20% of admail (Canwest News Service 2008), while the admail opt-out provisions of Canada Post do not apply to the remainder of the advertising which is delivered by private carriers. The admail WPA focuses on the amount of unwanted admail from private carriers, excluding advertising inserted into subscription newspapers. It is assumed that the current refusal percentage for this type of admail is identical to the admail opt-out percentage for Canada Post, since the sign objecting to admail that must be posted on one’s mailbox would presumably be seen and obeyed by the private carriers of admail as well. The amount of waste prevention depicted by the admail WPA is based on the difference between the 2% already opting out of admail, and the 67% of Canadians who find admail to be of no interest. This difference is equal to 10,007 tonnes, based on mass measurements of admail from private carriers, excluding advertising inserted into subscription newspapers, received in one week at a Toronto residence in Winter 2011 (see Section 5.3.4 and Appendix 4 for data collection procedure and analysis). 126

The assumed quantity of admail waste prevention undertaken in the waste prevention scenario (4.2 kg/capita) is comparable to the estimate of 5.7 kg/capita/yr by Salhofer et al. (2008) of the quantity of unwanted admail in Vienna, Austria. The per capita estimate by Salhofer et al. (2008) takes into account the results of a survey by Wassermann et al. (2004) which indicated that that less than half (47%) of households in Vienna, Austria wished to receive admail. Unlike Salhofer et al. (2008), the unwanted admail estimate in the waste prevention scenario excludes admail from postal carriers and advertising flyers within subscription newspapers. The disposable bag WPA exemplifies WPA type 2, and depicts the reuse of carry out disposable plastic bags. These bags have been targeted for waste prevention through the Ontario Plastic Bag Reduction Initiative, a voluntary agreement between government, industry and the Recycling Council of Ontario to reduce the number of plastic bags distributed in the province of Ontario, Canada by 50 per cent by 2012 (Ontario Plastic Bag Reduction Task Group 2008). Although much of the reduction would be a consequence of the substitution of reusable bags, an increase in the reuse level of disposable bags is also a possibility. Most disposable bags in Canada are already used two or more times (CPIA n.d.), frequently as garbage bags, and 90% of disposable bags are reused for various purposes, according to a poll commissioned by the Canadian Plastics Industry Association (Exchange Magazine 2006). In comparison, waste audits conducted by Stewardship Ontario in 2009 have indicated that 56.4% of plastic bags were reused as containers for garbage, source-separated organics or recyclables (Ontario Plastic Bag Reduction Task Group 2010). In light of these data, as well as the lack of published information on the current levels of disposable bag reuse in stores, this WPA uses an estimate of 20% of disposable plastic bags displaced through the reuse of these types of bags in stores. This reuse in stores needs not affect the levels of plastic bag reuse in the home, since the bags would presumably remain intact. This WPA is considered additional to the level of disposable bag reuse already taking place by consumers in Toronto stores in 2008. The newspaper WPA within the waste prevention scenario refers to the substitution of daily subscription newspapers with their online equivalent. While newsprint from newspapers has represented a substantial portion of residential waste, the 127

amounts generated have fallen steeply in recent years, reflected by annual contractions in the Ontario newspaper market of 12.5% in 2008 and 7.5% in 2009 (Canadian Newspaper Association 2010). Much of this reduced consumption could be attributable to the substitution of online newspaper articles for actual newspapers. Based on the mean of the 2008 and 2009 newspaper market contractions, an additional 10% reduction in the consumption of daily subscription newspapers from the 2008 levels is assumed, equal to 4,528 tonnes of waste prevention. In order for the waste prevention scenario to remain functionally equivalent to the 2008 reference, it is necessary to substitute the time that was used to read a newspaper with the equivalent amount of time spent on a computer reading newspaper articles downloaded from online and displayed on the computer monitor. The examples of WPA-4 and WPA-5 are based on the LCA by Cleary (unpublished / Chapter 4) addressing alternative packaging systems for wines and spirits consumed domestically (i.e., outside of commercial establishments) in the Toronto market, using 2008 as the reference year. For the refillable bottle WPA, domestic and bulk imported wines and spirits are assumed to be sold in refillable glass containers, representing 32.8% of the wine/spirit market (excluding wine in bag-in-box containers – see Section 5.3.2) by volume. The waste prevention examples for WPA-5 comprise the substitution of lightweight single use glass bottles, PET bottles and aseptic cartons, affecting the remaining 67.2% of the Toronto wine/spirit market by volume. Changes in the quantities and types of closures, capsules and labels used are addressed for these WPAs as well. The justifications for the assumptions used for the refillable bottle and lightweight packaging WPAs are described for the alternative packaging scenario in Chapter 4 and Appendix 3. Grasscycling is selected to exemplify WPA-7. The City of Toronto (2008) estimated that approximately 15% of yard waste was prevented due to grasscycling in 2008. Since the WasteMAP LCA model explicitly accounts for “additional WPAs,” only those net environmental impacts associated with grasscycling above this 15% level are addressed in this WPA. The Generally Accepted Principles (2003) that the City of Toronto uses to calculate municipal solid waste system flow assumes an absolute limit of 20% of yard waste prevention from grasscycling, a limit which is attained once a “three- 128

bag limit or lower for garbage, plus user pay” is implemented. This type of system was not in place in Toronto at the time. For the grasscycling WPA, the 20% level is adopted, which is equal to 4,028 tonnes of grasscycling in addition to the amount estimated in the 2008 reference scenario.

5.3.2 Functional units

Functional units are used in the WasteMAP LCA not only to ensure a consistent amount of residential waste managed in all scenarios, but also to maintain identical reference flows of functionally equivalent product services for the residents of the municipality. The primary functional unit (PFU) is the mass of the City of Toronto’s residential waste that was managed in 2008. This tonnage includes the mass of waste collected by the municipality and through the wine/spirit packaging deposit-return program, and the net amount prevented due to the WPAs. Thus, for the 2008 reference scenario, 830,253 tonnes of waste are collected by the municipality, and 13,439 tonnes are collected in the deposit-return program (843,692 tonnes in total). An identical amount is managed in the waste prevention scenario, and consists of 29,018 tonnes of waste prevention, 807,652 tonnes collected by the municipal government, and 7,023 tonnes of wine and spirit packaging waste collected in the deposit-return system. The secondary functional units (SFUs) quantify the performance of the product system(s) removed and added due to the WPAs (Cleary 2010 / Chapter 3). The product systems that provide the services measured by the SFUs are depicted as (1) the targeted product systems (TPSs) which are subtracted from the MSW management system; and (2) the alternate product systems (APSs), which are added to it (Cleary 2010 / Chapter 3). SFUs are not required for the examples of WPA types 1 and 7 because there are no product services consumed. The disposable bag WPA has an SFU of 1.19*108 bags. The SFU for the newspaper WPA quantifies the total amount of time reading newspaper articles (1.70*107 hours), either online or in printed format, affected by the product service substitution. For the examples of WPA types 4 and 5, the SFUs address the volume of packaged wines and spirits subjected to the packaging substitutions (3.27*107 litres). 129

5.3.3 System boundaries

Although the “cradle” of the 2008 reference scenario is set at the moment of residential waste generation, the system boundary of the waste management scenario incorporating WPAs must comprise both upstream and downstream components. The upstream processes are associated with the generation of the waste materials targeted for prevention, as well as the substitute product systems that generate less waste, if such substitutes are necessary. The downstream processes are those traditionally connected with MSW management, beginning with waste collection, and concluding with disposal or recovery through recycling and biological treatment (thermal treatment of residential waste is not undertaken in Toronto). Emissions associated with the production of equipment, infrastructure and energy are incorporated into the system boundary, although there are a few exceptions which are identified where necessary. Due to the incorporation of avoided burdens associated with the recycling of waste, the LCA system boundary is considered “cradle-to-cradle.” The displacement of conventionally- generated electricity and heat from energy supplied through the combustion of biogas is not credited to the waste management system since the treatment or prevention of residential waste is not the primary objective of the useful heat and electricity production process. These avoided burdens are not credited to the waste management system due to the potential of a perverse incentive to increase the production of decomposable waste for landfilling and biological treatment. To facilitate the interpretation of results, the system boundary of the 2008 reference scenario is limited to that which is traditionally used for an LCA of a residential waste management system with at-curb collection. Therefore, it excludes waste management activities to divert non-blue box recyclables (such as iron scrap), as well as hazardous, bulky, electronic and special wastes, totalling 7,809 tonnes of residential waste, although these forms of waste are included when picked up at the curb and landfilled. For the same reason, the 2008 reference scenario also excludes WPAs already implemented (i.e., not considered “additional”), such as backyard composting. The wine and spirit container wastes collected through the deposit-return/stewardship program are 130

included in the 2008 reference scenario because they will be affected by the WPAs in the waste prevention scenario. The system expansion depicted in the WasteMAP LCA model, which permits functionally equivalent scenario comparisons between the reference and alternative scenarios, applies to the waste prevention scenario (see Section 5.3.1). This scenario incorporates “additional” WPAs, those that are not implemented in the 2008 reference scenario. The system expansion for the waste prevention scenario also encompasses alternative waste collection processes when the incorporation of the WPAs results in a change to these processes. For example, the mass of material collected through the province’s deposit-return system for wine/spirit containers is reduced if the conventional containers are replaced with those of a lighter weight. The upstream system boundaries vary with the WPA. For the examples of WPA types 1 (admail) and 3 (newspapers), the upstream system boundary includes the processes related to raw material acquisition, manufacture and transport of newsprint. The newspaper WPA also addresses the life cycle processes for the production and use of computers, including data downloading from the internet. The disposable bag WPA comprises the raw material extraction and processing of high density polyethylene (HDPE) into bags, and the transportation of these bags to market. The upstream system boundaries for the examples of WPA types 4 and 5 incorporate the LCA processes related to (1) the extraction of the raw materials required for each container, and the secondary packaging (capsules, closures, labels) materials; (2) the transportation of the raw materials to processing facilities; (3) raw material processing; (4) the transportation of the processed materials to the container manufacturer; (5) the manufacture of the container; (6) the transportation of the containers to the wine and spirit packager; and (7) the transportation of the filled containers from the packaging facility to the City of Toronto. Only those processes associated with the product systems targeted for removal (e.g., conventional wine and spirit packaging) and the substituted product systems (e.g., alternative wine and spirit packaging) are included within the upstream components of the waste prevention scenario. Following the WasteMAP model, upstream emissions are 131

excluded from the 2008 reference scenario system boundary because no waste prevention is deemed to be undertaken (Cleary 2010 / Chapter 3). The downstream phase of residential waste management begins with collection of the waste and includes the transportation, sorting, treatment and disposal processes. When addressing refillable wine and spirit bottles, the system boundary encompasses the transport of empty used glass bottles to and from the retailer and the cleaning and refilling facility, as well as the cleaning process itself. By default, the results using WasteMAP incorporate the avoided burdens of recycling (Cleary 2010 / Chapter 3). All net benefits from the recycling of collected residential waste are allocated to the waste management system. However, avoided environmental burdens associated with upstream recycling processes (e.g., from the recycling of glass containers broken during the filling process) are excluded from the WasteMAP system boundary because they are considered a component of industrial waste management.

5.3.4 Data sources

This LCA is undertaken using SimaPro 7.2 LCA software using site-specific input data collected by the author. These data are used as inputs for those relevant LCA unit processes (see Appendix 1 for definition), such as glass production, either supplied in the EcoInvent American (US-EI) database, the Franklin USA 98 database (to a much lesser extent), or defined by the author and based on published or primary data. Each LCA scenario comprises a unique combination of unit processes that represent those processes within the LCA system boundary. The results from the evaluation of each LCA scenario are compared using the following life cycle impact assessment (LCIA) methods: (1) the American TRACI 2 (version 3.01) method (Bare et al. 2003); (2) the European Impact 2002+ (version 2.05) method (Jolliet et al. 2003); and (3) the European ReCiPe (version 1.02) hierarchist perspective (Goedkoop et al. 2009) LCIA method. The LCA results are evaluated using a variety of LCIA methods in order to supply a range of impact assessment interpretations of identical LCA process inputs. The TRACI 2 method is selected because its impact categories and methodology were designed to be 132

applicable to geographical circumstances in the United States (Bare et al. 2003), which has a much closer proximity to Canada than the continent of Europe. However, the low number of midpoint impact categories (nine) and lack of endpoint impact indicators limit the detail of the results analysis. In comparison, the Impact 2002+ has 15 midpoint impact categories, whereas ReCiPe has 18. The latter method is one of the most recent comprehensive LCIA methods available, and includes endpoint indicators. Impact 2002+, another prominent LCIA method with endpoint indicators, is also selected to provide a contrast with the endpoint results using ReCiPe. Primary and secondary data sources for the upstream component of this LCA include government and industry statistics, the published scientific literature, equipment manufacturers, as well as field research undertaken by the author. For the downstream component, data have been acquired from Solid Waste Management Services (SWMS) of the City of Toronto (2010a, 2010b, 2008), Stewardship Ontario (2006a, 2006b), and Waste Diversion Ontario’s (WDO) Municipal Datacall (2008). Although some of the required data are publicly available through published documents, much unpublished data have been obtained directly from the City of Toronto’s SWMS division. When site- specific data sources for unit processes were unavailable, the most appropriate and justifiable generic data supplied in LCA databases are used (see Appendix 4 for a list and description of the selected LCA processes). There are three main sources of data to estimate the amount, composition and treatment location of the collected waste. These sources include the published results of the City of Toronto’s estimates of the residential waste collected in 2008 (including residual and recyclable waste, as well as the waste subject to biological treatment), residential waste audits, and detailed tonnage data collected by WDO on the amount of recyclables collected by the City and shipped to various recyclers. Since not all of the profile datasets used for this comparative LCA apply identical definitions and assumptions for their parameters, data normalization is sometimes required in order to ensure that all of the data are consistent. Instances of data normalization are identified in Appendix 4. Field research was required for a plausible estimate to represent the amount of admail waste prevention that can be undertaken in the City of Toronto, without affecting 133

the quantities delivered to those who would want to receive it. It was first necessary to classify the different types of admail. Unaddressed advertising mail was delivered to 997,000 households in Toronto in 2008 (WDO 2008) through the postal system, and through private carriers. Admail from private carriers can be delivered in concert with subscription and community newspapers, as well as on its own. Only admail (mainly flyers) delivered separately or with free community newspapers by private carriers is taken into account for this WPA. The field research to determine the average mass of this type of admail delivered to Toronto households is based on the mass of admail received at a Toronto household over one week in Winter 2011 (7-13 February 2011). The estimate calculated from the field research is likely conservative, especially since the time period studied excludes admail peaks associated with major shopping periods such as Christmas/“Boxing Week.” Since subscription newspapers by their nature are wanted by the subscribing households, the advertising inserts associated with these newspapers are excluded from the admail WPA. For the newspaper WPA, field research was undertaken to estimate the average mass of daily subscription newspapers read in the City of Toronto. Statistics from the Canadian Newspaper Association (2009) indicate that the City of Toronto supports four paid daily English language newspapers, including the Toronto Star (35.1% of papers sold in the average week), The Globe and Mail (29.8%), the National Post (17.7%), and the Toronto Sun (17.4%). In order to produce an estimate of the average mass of these newspapers, a sample of each newspaper was weighed from Monday, January 17th to January 23rd 2011 using a digital kitchen scale. The mass of each newspaper was then multiplied by the total number of newspapers of the particular publication that were sold on the given day of the week in 2008. It is assumed that the market for these newspapers is the Greater Toronto Area (Census Metropolitan Area), which has approximately 2.2 times the population of the City of Toronto (City of Toronto 2006, Statistics Canada 2008a, 2006). The detailed results of the field work for the admail and newspaper WPAs are listed in Appendix 4, along with additional information on the procedure used in undertaking the project. Since this field research was pursued in 2011, the average mass 134

of the advertising mail and newspapers might have changed since 2008, although this change is assumed to be negligible.

5.4 LCA input profile: Mass balance

Using the WasteMAP LCA model, cumulative net residential waste generation (RNET) for the waste prevention scenario is calculated through the following equation (Equation 1 from Cleary 2010 / Chapter 3), assuming a total of “n” WPAs (for this scenario there are six WPAs):

n n RNET =RAPSWPA – RTPSWPA WPA=1 WPA=1

where RAPS is the residential waste generation potentially added to the MSW treatment system due to product substitutions associated with WPAs, and RTPS is the residential waste subtracted from the MSW treatment system. Net residential waste prevention requires RNET to be less than zero. Therefore, for example, the masses of the wine and spirit container and secondary packaging waste needed to be less than the amount calculated for the 2008 reference.

5.4.1 2008 reference scenario

When applying the WasteMAP model for waste management scenario comparisons, all upstream LCA processes are excluded from the system boundary of the reference scenario which lacks the selected WPAs. Figure 24 illustrates the waste flows taken into account under this 2008 reference scenario.

135

Reference Scenario (9) (4) Residential Waste Transfer Station Composting Facility (1) Waste for (2) Municipal (5) Collection Anaerobic Digester (6) (10)

(3) (7) Landfill

Material Recovery Facility Wine/Spirit (8) (11) Container Deposit- Return System (12) Recycling Facility

Figure 24 Flows of waste under the 2008 reference scenario for the life cycle assessment of the City of Toronto’s residential waste management system. The magnitudes of the waste flows associated with each waste flow identifier are defined in Table 14.

Data from the City of Toronto (2010a, 2010b, 2008), Waste Diversion Ontario (WDO) (2008) and the Stewardship Ontario (2006a, 2006b) waste audit program are used to estimate the quantities of residential waste generated, collected and treated. The overall quantities collected and treated are supplied by the City. The material composition of the residual waste (garbage) for landfilling is estimated by extrapolating data from published waste audits of single family and multiple unit dwellings in Toronto in 2005 and 2006. WDO data are used to estimate the material composition of the recyclable wastes collected. The procedures used to estimate the material composition of the residual and recyclable wastes are described in detail in Appendix 4. Excluding the residual waste from the recycling processors, 494,539 tonnes of Toronto’s waste were sent to the Carleton Farms landfill in the state of Michigan (93.2% of total waste) and the Green Lane landfill located between St. Thomas and London, Ontario (6.8% of total waste) in 2008 (City of Toronto 2010a). During that year, 158,747 tonnes of residential waste were supplied to recycling facilities, 94,201 tonnes were transported to source-separated organics treatment facilities, and 82,766 tonnes of yard waste went to composting facilities (City of Toronto 2008). In 2008, Toronto’s material recovery facilities (MRFs) produced a 17% residue from the recycling materials collected at curbside (also known as “blue box” recyclables) and received for processing, all of which was landfilled (City of Toronto 2010a). Using this 17% residue estimate, the amount collected for recycling was 32,514 tonnes greater 136

than the 158,747 tonnes estimated by the City, which was the 2008 output from the MRFs. Waste Diversion Ontario (2008) estimates that 117,751 tonnes of source separated organic (SSO) waste were collected by Toronto and treated in 2008. In contrast, the City of Toronto estimates a much lower 94,201 tonnes. This difference is the 20% non- organic residue included with the collected SSO waste. Table 13 identifies the amount of residential waste generated in the City of Toronto in 2008, by waste material type.

Table 13 Residential waste generation in the City of Toronto, by material composition, 2008

Composition of Residential Waste Mass (tonnes) % of Total Residual waste to landfill 494,539 58.6% [Including losses at recycling processors] [498,879] Deposit-return system for wine/spirit containers 13,439 1.6% Green Bin, Source Separated Organics (SSO) 94,201 11.2% Leaf and Yardwaste, Christmas Trees 82,766 9.8% Recycling (output of blue box recyclables after sorting at MRFs) 158,747 18.8% [Accounting for losses at recycling processors] [154,407] Mixed paper 88,687 Cardboard 32,970 Glass 17,773 Commingled 7,226 Steel 4,762 PET 3,028 HDPE 2,092 Aluminum 1,091 Poly-coat 787 Plastic (tubs) 280 Polystyrene 41 Plastic film 10 Sources: City of Toronto (2008), with recycled material estimates based on Waste Diversion Ontario (2008) figures normalized to City of Toronto residential recyclables estimate. The amount of wine/spirit packaging waste collected through the deposit-return system was calculated in Cleary (unpublished / Chapter 4).

5.4.2 Waste prevention scenario

In order to allow the incorporation of WPAs into the LCA of Toronto’s waste management system, it was necessary to expand the system boundary of the waste prevention scenario relative to that of the reference. In contrast to Figure 24, Figure 25 also illustrates the altered waste flows resulting from the targeted and alternate product systems, including the net negative flows from the container deposit-refund system. The flows of waste under each waste management scenario are depicted in Table 14. 137

Waste Prevention Scenario (1) (9) (4) Net Waste Prevention Residential Waste Transfer Station Composting Facility (13) Waste for (2) Municipal (5) Collection Anaerobic Digester Targeted Product Systems [Products eliminated from (14) (6) (10) waste stream] (3) (7) Landfill

Alternate Product Systems Material Recovery Facility [Goods which supply substitute (15) (8) (11) product services added to the waste stream] Recycling Facility Wine/Spirit Container Deposit- (12) (16) Return System

Figure 25 Flows of waste under the waste prevention scenario for the life cycle assessment of the City of Toronto’s residential waste management system. Dashed arrows represent negative flows of waste. The magnitudes of the waste flows associated with each waste flow identifier are defined in Table 14.

Table 14 Waste flows of each waste management scenario

Stage of life cycle Mass of waste (tonnes)2 Waste Description 2008 reference scenario Waste prevention flow scenario identifier1 1 Residential waste for municipal collection 830,253 807,652 2 Waste to waste transfer stations 830,253 807,652 3 Waste to material recovery facilities 191,261 178,368 4 Waste from transfer stations to composting facilities 82,766 78,738 5 Waste from transfer stations to anaerobic digestion 117,751 117,751 facilities 6 Waste from transfer stations to landfills 438,475 432,795 7 Waste from MRF to landfills 32,514 30,323 8 Waste from MRF to recycling facilities 158,747 148,045 9 Waste (digestates) from anaerobic digestion 94,201 94,201 facilities to composting facilities 10 Residual waste from anaerobic digestion facilities 23,550 23,550 to landfills 11 Residual waste from recycling facilities to landfills 4,340 4,113 12 Waste from wine/spirit container deposit-return 13,439 7,023 system sent to recycling facilities 13 Prevented waste that would have been collected by 0 -26,065 the municipality for treatment 14 Additional waste (for municipal collection) from 0 3,463 goods supplying substitute product services 15 Prevented waste that would have been collected 0 -13,439 through the deposit-return system for treatment 16 Additional waste (for deposit-return system) from 0 7,023 goods supplying substitute product services 1 Waste flow identifiers refer to the numbers within parentheses next to each waste flow arrow from Figures 24 and 25. 2 Although for calculation purposes in the model the waste prevention and collection estimates are rounded to the tonne, the specific numbers of digits associated with all of these estimates are not intended to indicate a particular level of certainty (uncertainty is addressed in Section 5.8).

138

Since the WPAs in the waste prevention scenario affect the amount and composition of waste to be treated, it is assumed that the proportions of each waste stream that were landfilled, recycled and biologically treated, under the waste prevention scenario, are identical to the proportions in the 2008 reference year.

5.5 LCA input profile: Unit processes

The unit processes within the system boundaries of the residential waste management scenarios are mostly adaptations of those derived from the US EcoInvent (US-EI) database. For cases in which the US-EI and other databases associated with SimaPro 7.2 lacked particular LCA processes needed for this LCA, it was necessary to collect the relevant data to define these processes manually. These data were acquired from published peer-reviewed literature, equipment manufacturers and operators. In some cases, an LCA unit process supplied by a database was altered in order to allow it to be more representative of a specific component of the waste management system. Author-defined LCA processes include the following: (1) aseptic carton manufacture; (2) bottle washing; (3) glass production and recycling; (4) the sorting of recyclable materials at material recovery facilities (MRFs); (5) HDPE recycling; (6) PET recycling; and (7) anaerobic digestion of source-separated organic waste. A comprehensive list and description of those unit processes selected for the LCA scenarios is available in Appendix 3 and 4. Although the geographic origin of the EcoInvent data is mainly European, this database is one of the most comprehensive and commonly used for published LCAs of systems within a wide variety of geographic contexts. The geographic locations of the technologies used for the LCA processes are wide-ranging. Implicit to the LCA is the assumption that the technological differences between the various countries are negligible. However, the differences in the electricity production mixes are less likely to be negligible, and are taken into account in the LCA. Where indicated in the list of unit processes used in the LCA scenarios (see Appendix 4), the electricity mixes of most unit processes from the LCA databases are replaced by the author with ones more representative of the geographical context of the modelled processes. However, the 139

electricity mixes for any electricity inputs for the process, material or capital inputs to the unit processes remain the defaults for the original unit process derived from the LCA databases. For those processes within the system boundaries of the LCA which required electricity, the environmental emissions are based upon the average electricity supply mix from the 2008 national electricity transmission grid of the country where the process took place. Those processes occurring in Canada use the 2008 Ontario electricity supply mix. Electricity supply mixes are derived from International Energy Agency (IEA) 2008 statistics as well as those from Ontario's Independent Electricity System Operator (2009). Appendix 4 lists the electricity supply mixes and describes in detail the procedure used to incorporate these electricity mixes into the unit processes.

5.5.1 Upstream processes

For the waste prevention scenario, the impacts from the net changes to the upstream processes associated with the production, transportation and use of those products affected by the WPAs are within the system boundary. The quantities of the upstream unit processes used as inputs for the waste prevention scenario depicted using the SimaPro 7.2 program are listed in Appendix 4.

5.5.1.1 Reduced generation of unaddressed advertising mail (“junk mail”)

For this WPA, it is assumed that the advertising mail is composed of newsprint, based on the field research which indicated that newsprint comprised approximately 88% of the measured admail mass, with the remainder found to be “glossy” paper. The average transport distance from the manufacturer to Toronto is assumed to be 541 km, with 75% shipped by rail and the remainder shipped by truck (see Appendix 4 for justifications). A conservative estimate of 40% for the recycled content of the newsprint is selected, based on the minimum recycled content of newsprint used for the Toronto Star (Toronto Star n.d.), the daily newspaper with the largest circulation in Canada 140

(Canadian Newspaper Association 2009). No alternate product system is required for this WPA example.

5.5.1.2 Reuse of disposable plastic shopping bags

For the disposable bag WPA, it is assumed that the disposable plastic shopping bags are composed of HDPE plastic with 15% recycled content, based on ICF Consulting’s (2005) estimate for HDPE in Canada. The reused bags are assumed to have an average mass of 7 grams per bag, based on Verghese et al. (2009), and in keeping with the Canadian Plastics Industry Association’s estimate of 6-7 grams per bag (CPIA n.d.). The Ontario Plastic Bag Reduction Task Group (2010) claimed that three billion disposable plastic bags were distributed in Ontario in 2008. Accounting for the percentage of Ontario’s population living in Toronto (19.76%), the disposable bag WPA is estimated to prevent 1,216 tonnes of waste. There are no published data on the mean transportation distance of disposable plastic bags from the manufacturer to the location of consumption in Toronto. Therefore, the relatively short trucking distance of 100 km distance is used because the Greater Toronto Area is the largest manufacturing centre of Canada and it would be unlikely that the 426 km transport estimate used for Canada by ICF Consulting (2005) in its GHG LCA of Canadian industries would be valid for this LCA.

5.5.1.3 Substitution of newspaper articles available online for those printed on newsprint

Although the process input assumptions for the newspaper WPA are identical to those for the admail WPA (i.e., transportation and recycled content), the quantity of waste prevention undertaken is somewhat less. Unlike the admail WPA, a functionally- equivalent substitute, in this case, a substitute for the time spent reading the newspaper, is required. This newspaper WPA uses the estimate from an analysis by Scarborough Research and Newspaper National Network LP (2010) that the top 25 newspapers in the United States had an average readership of 3.18 readers per copy in 2008. In order to 141

calculate the amount of electricity used to download and display newspaper content online, the WPA also adopts the estimate that the average person who reads a newspaper in print or online reads it for 20 minutes per day, the average of the 30-minute estimate by Moberg et al. (2010) and the 10-minute estimate of Hischier and Reichart (2003). It is assumed that the online articles require 5 MB of downloaded content per reader each day, an estimate used for e-paper newspapers in the LCA by Moberg et al. (2010). A unit process depicting the kWh of electricity used for the transmission of downloaded content is incorporated into this WPA. An electricity intensity of 7 kWh/GB downloaded is adopted, based on the estimate by Weber et al. (2010) for 2008. The environmental burdens from the production and disposal of additional computer equipment are taken into account. The impacts from the production and maintenance of internet servers and networks are excluded, and assumed negligible relative to those from the use of electricity for the computer hardware and internet downloading.

5.5.1.4 Substitution of refillable and lightweight wine and spirit containers for conventional containers

All assumptions pertaining to the production and transportation of the containers and secondary materials are derived from the product LCA of wine and spirit containers in Chapter 4. The main data inputs for the unit processes are based on statistics from the Liquor Control Board of Ontario (2008), Statistics Canada (2008a, 2008b), the Association of Canadian Distillers (2008) and the Canadian Vintners Association (2008). For the refillable bottle WPA, the refillable containers for domestic wines and spirits are used an average of 15 times, based on the estimate generated by The Beer Store (TBS 2009). A 100 km average transport distance between the refillable container collection centres in Toronto and the bottlewashing/refilling facilities is assumed for this WPA, based on the product LCA by Cleary (unpublished / Chapter 4). Transportation data and assumptions for non-refillable containers are derived from Statistics Canada (2008b) import statistics and questionnaires to wineries and distilleries (Cleary unpublished / Chapter 4).

142

5.5.1.5 Grasscycling

Within the LCA, the grasscycling WPA affects only the downstream component by reducing the quantity of yard waste collected and treated. It is assumed that no new equipment, such as a mulcher, would need to be manufactured and used in order to increase the level of grasscycling by 5%, or 4,028 tonnes.

5.5.2 Downstream processes

Downstream processes are taken into account in both the 2008 reference and waste prevention scenarios. Table A4.9 in Appendix 4 provides the quantities of the downstream unit processes used as inputs for the different LCA scenarios. For those processes which require electricity, the environmental emissions are calculated using the average electricity supply mix for Ontario in 2008. The sole exception is for those processes associated with the landfilling of Toronto’s residential waste, almost all of which was sent to the state of Michigan, where the US electricity mix, as defined in the US-EI database, was utilized. Since 2007, a portion of the generated residential waste, namely beer, wine and spirit containers, has been collected at The Beer Store (TBS) which operates the Ontario beverage container deposit-return system. The net environmental burdens associated with this system are excluded from the reference scenario in order to streamline the LCA, limiting the reference to the treatment of those residential wastes collected by the municipality. However, since the WPAs pertaining to wine and spirit packaging would not only affect the quantities of waste collected by the municipality, but also the amounts collected through The Beer Store, it is important to account for the net changes to the TBS collection system in the waste prevention scenario.

5.5.2.1 Residential waste collection and transportation

In the City of Toronto, the collection of source separated organics occurs once per week, while recyclables and residuals (garbage) are collected in alternating weeks. The 143 environmental emission estimates for the downstream transportation component are based upon the fuel use and the average load of the waste collection and transfer vehicles, as well as the mean distances that the wastes travelled to the landfill, sorting, transfer, recycling and organics treatment facilities in 2008. These estimates also take into account the processes associated with the production of the waste collection and transfer vehicles. Unpublished data acquired from the City of Toronto (2010a) to estimate the waste transportation emissions include the following: (1) the locations (i.e., the city) where each type of residential waste material was sent in 2008; (2) the approximate percentages of each type of residential waste stream sent to each treatment/disposal facility in 2008; and (3) the average distance (19.75 km – City of Toronto 2010a) that Toronto’s waste collection vehicles transported waste. The procedures, assumptions and input data used to estimate the environmental burdens from residential waste collection and transportation are described in Appendix 4.

5.5.2.2 Sorting

All collected residential waste in Toronto is transported either to waste transfer facilities or to material recovery facilities (MRFs). At both of these types of facilities, the waste is transferred to vehicles with larger hauling capacities. At the MRFs, the recyclables, which were collected in a commingled state, are sorted before sending the material onward to the recycling processors. Waste allocated to landfill (garbage) is not sorted. The issue of removing residuals from the green bin contents at the organics treatment facilities is addressed in Section 5.5.2.3 (Biological Treatment). Residential waste sent to landfill also originates from waste treatment facilities, including the contaminants from the MRFs, as well as from the biological treatment and recycling facilities. The environmental emissions from the sorting of mixed recyclables at MRFs are estimated through an author-defined unit process which includes MRF energy inputs based on Franklin Associates (2010), and the material (non-waste) and capital inputs from the US-EI process entitled “waste paper sorted, for further treatment.” A detailed description of this unit process is available in Appendix 4. 144

5.5.2.3 Biological treatment

The City of Toronto’s organic waste is classified into yard waste (including Christmas trees) and source separated organic (SSO) waste streams. The former is composted while the latter is subject to both anaerobic digestion and composting once the waste is processed to remove contaminants. The resulting compost from both types of organic waste is then used for farms and parks (City of Toronto 2010b), as well as for private gardens (City of Toronto 2011). The existing unit process for composting in the US-EI database is included in the LCA to depict the treatment of yard waste and SSO digestates. I have designed an LCA process using SimaPro software in order to estimate the emissions associated with the processing and anaerobic digestion of the SSO waste. Inputs for this process are based upon the 2008 operating records for Toronto’s Dufferin Organic Processing Facility (DOPF). The SSO treatment process undertaken at the Dufferin facility, which received approximately 40% of the City’s SSO waste in 2008 (City of Toronto 2010a), is assumed to be representative of that taking place at the other facilities that processed Toronto’s SSO waste. The DOPF operating records indicated that, for every tonne of SSO material entering the facility, approximately 50.3 kWh of electricity and 7.4 m³ of natural gas were used. Fresh water use at the facility was approximately 0.61 m³ (610 litres) per tonne of SSO waste input, while an average of 0.94 m³ of effluent was generated (City of Toronto 2010a). Biogas production for flaring ranged from 100-120 m³/tonne of SSO processed, although the expected yield is approximately 120, which was assumed in the SSO waste treatment unit process. In 2008, most (67.5%) of the SSO digested solids, equivalent to 32-35% of SSO mass inputs (33.5% assumed), were shipped 135 km to Arthur, Ontario for composting, with the remainder travelling 120 km to a composting facility in Thorold, Ontario (City of Toronto 2010a). The proportion of the SSO inputs ending up as solid processing residue in 2008 was approximately 20%, based on the difference between estimates provided by Waste Diversion Ontario (2008) for SSO waste curbside tonnage collected (117,751 tonnes) and the City of Toronto (2008) for actual SSO waste diversion (94,201 tonnes). Appendix 4 provides a detailed description of the 145

author-defined unit process for the anaerobic digestion of source-separated organic waste material. The 37.5% reprocessing efficiency estimate of Rigamonti et al. (2009) for the composting of organic waste is used to calculate the mass of compost generated from the yard wastes. In contrast, additional material losses from composting the SSO waste digestates are assumed negligible because the anaerobic digestion process already results in the decomposition of most of the material content, while stabilizing much of the remaining mass. Avoided emissions resulting from the use of the compost products of organic waste were excluded from the system boundary because it is questionable that, had the compost not been produced, the consumers would have purchased a substitute material.

5.5.2.4 Recycling

Recycling is a two-step process. The first step is the sorting of the commingled recyclable material inputs, while the second is the reprocessing of these materials. The City of Toronto collects numerous forms of recyclable waste through its “blue box” program. These materials are then transported to one of two MRFs to remove contaminants and separate each type of recyclable material in order to allow relatively pure feedstocks to be sent to each recycling processor. For both scenarios, recycling is addressed using the “avoided burden” allocation, which takes into account the unit processes to produce new material with recycled content, and subtracts the avoided processes to produce the same types of material using only virgin content. Although many of the recycling processes incorporated into this LCA originate from LCA databases, I have also put together unit processes in SimaPro using data published in scientific and industry literature (recycled HDPE, recycled PET, recycled glass). Appendices 3 and 4 have lists of the unit processes depicting recycling, and the quantities of each process that are incorporated into the LCA scenarios. There are significant losses of recyclable material inputs, as contaminants and processing losses, at the re-processors. Therefore, it is inappropriate to assume a mass substitution ratio of 1:1 for the recyclable material inputs and the material outputs. The 146

reprocessing efficiencies for various types of recyclable material are taken into account in calculating the burdens from the recycling processes, and the avoided burdens from the displacement of virgin materials. Table 15 displays a list of the waste stream diversion and reprocessing efficiencies used in the LCA calculations, and identifies the avoided product for each recycled waste stream. It is assumed that these diversion and efficiency percentages do not vary under each scenario. As waste from recycling facilities is excluded from the municipal estimate of landfilled waste (City of Toronto 2008), this waste stream, which can be considered IC&I waste, is also omitted as an input to landfills for this LCA.

Table 15 Percent diversion of each residential waste stream and reprocessing efficiency at recyclers and biological treatment facilities

Waste material % diversion of waste stream from landfill to Reprocessing efficiency (%)2 stream municipal recycling or biological treatment1 Aluminum 26.7 (displacing virgin aluminum) 97.1% (US-EI estimate in aluminum recycling) Cardboard 40.2 (displacing cardboard) 97.1% (US-EI estimate in corrugated board from recycling fibre) Glass 57.2 (displacing bottle glass) 100% (Rigamonti et al. 2009) HDPE 50.1 (displacing virgin HDPE granulate) 92.7% (HDPE – Franklin Associates 2010) Organic waste 52.5 (no displacement assumed) 37.5 (compost product - Rigamonti et al. 2009) Paper 66.0 (displacing newsprint) 97.3% (US-EI estimate in newsprint production) 4 PET 40.8 (displacing virgin PET granulate) 80% (PET – Franklin Associates 2010) Plastic film 0.0 (displacing virgin LDPE film) 91.0% (Franklin USA 98 estimate in recycled LDPE unit process) Polycoat 19.2 (displacing cardboard) 97.1% (US-EI estimate in corrugated board from containers recycling fibre) 3 Polystyrene 0.8 (displacing virgin polystyrene) 91.0% (Franklin USA 98 estimate in recycled PS unit process) Steel 55.7 (displacing pig iron) 100% (US-EI estimate for processing of iron scrap) 1 Estimates based upon City of Toronto and Waste Diversion Ontario statistics. Diversion figures exclude collection through the container deposit-return program. 2 Defined as the percent of waste inputs that are incorporated into the final product. 3 This estimate only pertains to the paperboard component of polycoat containers. Polycoat containers are assumed to comprise 76.6% paperboard, based on the average for 1 L and 1.5 L aseptic cartons. 4 Includes the losses from the disposal of MSW and ash from de-inking sludge.

For those WPAs which affected the quantities of waste collected through the deposit-return system, the container return rates remain the same as those measured in the 2007-2008 fiscal year (i.e., 29% for aseptic cartons, 34% for PET bottles, and 69% for glass bottles) (TBS 2008). The wine and spirit packaging WPAs would mostly affect the quantity of glass in the recycling stream. Although some glass is not recycled back into its original product type (i.e., a glass container), a dearth of published data on the recycling of glass into fibreglass and other products necessitates the assumption of a 147

closed-loop “bottle-to-bottle” recycling system for the City of Toronto’s waste glass. Forms of open-loop recycling for glass may include using glass as an aggregate substitute, a sand substitute, an abrasive material, a filter material; and as a raw material in the fibreglass industry (Unical 2008, Toronto City Council 2002). Unlike what occurs in municipal waste collection, the Province of Ontario’s deposit-refund system for wine and spirit containers allows glass to be separated according to colour, and recycled into glass bottles. Non-refillable clear glass bottles collected through this system are remanufactured into new bottles, 50% of coloured glass bottles are recycled into fibreglass, with the remainder used for manufacturing new coloured containers (TBS 2009). The majority of recycled PET bottles are used for plastic strapping and felted automotive materials, while aseptic cartons are combined with old corrugated cardboard and boxboard for recycling (TBS 2009).

5.5.2.5 Landfilling

In 2008, Toronto’s residential waste was disposed of at a sanitary landfill in Carleton Farms, Michigan, and at the Green Lane landfill near St. Thomas, Ontario. Approximately 93.8% of the 494,539 tonnes of residual waste was sent to the former, which is more than double the distance from Toronto than the Green Lane landfill (City of Toronto 2010a). This quantity, provided by the City of Toronto, includes the contaminants from both the MRFs and the biological treatment facilities, but excludes the processing wastes from recycling. Both landfills have equipment to collect landfill gas, composed of approximately 50% methane, emitted from waste decomposition. The collected gas is burned for electricity generation. The US-EI unit process is used to depict landfilling and assumes that 53% of landfill gas is captured for flaring or useful heat and electricity production (Doka 2009).

5.5.2.6 Reuse of wine and spirit containers

Since the refilling of reusable wine and spirit containers currently does not take place in Ontario, 200 km is used as a plausible round-trip distance to transport empty 148

glass bottles between the retailer (The Beer Store) in Toronto and the bottlewashing/refilling facilities (i.e., approximately the round-trip distance between Toronto and the Niagara wine region between Stoney Creek and Niagara on the Lake). This estimate appears reasonable in light of the suggestion by Valiante (2007) that the typical round trip for a refillable wine bottle in Ontario “might be in the order of just 300 km.” As this case study focuses on the City of Toronto, the largest market in Ontario, it is likely that the distance would be substantially less than Valiante’s (2007) estimate for the entire province. Recent technical data on bottle washing using modern equipment for 1000 ml glass wine containers were recently acquired from a bottle washing equipment manufacturer in order to facilitate the design the bottle washing LCA process within SimaPro 7.2.

5.6 Life cycle impact assessment (LCIA) results

Life cycle impact assessment is used to evaluate the magnitude and significance of the potential environmental impacts which result from the processes within the system boundary of the LCA (ISO 2006). Three prominent LCIA methods – ReCiPe v1.02, Impact 2002+ v2.05, and TRACI 2 v.3.01 – are used to evaluate and compare the results of the different LCA scenarios. ReCiPe uses three perspectives to group similar types of assumptions and choices in the impact characterisation models: hierarchist, individualist and egalitarian. ReCiPe’s hierarchical perspective is employed because it is “based on the most common policy principles with regards to time-frame and other issues” (Goedkoop et al. 2009). Thus, the results from ReCiPe take into account environmental persistence of the emitted substances. Similarly, TRACI 2 also uses limited time horizons. Impact 2002+, unlike ReCiPe and TRACI 2, uses infinite horizons (independent of persistence of the environmental emissions) to incorporate the long-term effects of each environmental burden over time (EC JRC 2010). Major differences between the selected LCIA methods are discussed in the documentation associated with each method, as well as in the European Commission Joint Research Centre (2010) comparative analysis of prominent LCIA methods. Other than time horizons, the selected LCIA methods differ in the number of modelled 149

substances incorporated into each method, the number and types of impact categories included, the units used to depict some of the impacts, and the characterisation factors used in defining the midpoint and endpoint impact categories. For example, TRACI 2 accounts for impacts from its acidification category as moles of H+ equivalents (Bare et al. 2003), while Impact 2002+ has separate impact categories for terrestrial and aquatic

acidification, both of which are measured in kg of SO2 equivalents (Jolliet et al. 2003). The differences between the selected LCIA methods can make it problematic to compare absolute figures from the impact categories of each respective method. Nevertheless, there are numerous areas of similarity, especially in terms of the types of impacts, such as climate change, and the units used in depicting the impact magnitudes. Comparisons of the results are undertaken using both midpoint and endpoint level impact category indicators (Sections 5.6.1 and 5.6.2). Midpoint level indicators describe the magnitudes of various types of environmental changes, such as water acidification. In contrast, endpoint level indicators depict the damages resulting from environmental changes as reductions in environmental quality, resource availability, and human health (Jolliet et al. 2003). The additional step of incorporating data from the various midpoint level impact categories into the endpoint level (damage) impact introduces further uncertainty to the results, the degree to which is described in the documentation associated with each LCIA method. It should be noted that, although ISO 14044 requires LCA scenario result comparisons to link to category endpoints (ISO 2006), the TRACI 2 method does not supply endpoint level indicators. Since environmental emissions calculated using the LCA tool tend to be summed up without reference to the particular location of the emission, it is difficult for one to claim whether or not particular impact thresholds (e.g., the level of aquatic acidification sufficient to kill most algae) have been exceeded. The design of those LCIA methods which include endpoint/damage indicators addresses this issue, imperfectly, by assuming the average conditions for the identified geography of each method (i.e., United States using TRACI 2; Europe using Impact 2002+ and ReCiPe). The findings presented below indicate that the waste prevention scenario was superior to the 2008 reference for the midpoint and endpoint level indicators included in the selected LCIA methods. The only exceptions, the land occupation midpoint impacts, 150

showed very small increases (<2% relative to the 2008 reference scenario), although the avoided burdens more than compensated for these increases. These exceptions result from the slight increase in forest harvesting from the reduction in newsprint waste entering the recycling system that is associated with waste prevention in the admail and newspaper WPAs.

5.6.1 Midpoint level comparisons

Most midpoint level indicators quantify the LCA results using units unique to

each impact category (e.g., CO2 equivalents, SO2 equivalents). Therefore, the indicator values from different midpoint impact categories are not directly comparable with one another on an equivalent basis. Figures 26, 27, and 28 display the net percent change in the midpoint level impacts from the 2008 reference scenario to the waste prevention scenario. Impacts and avoided burdens are calculated by summing the positive and negative amounts of each contribution, by substance, to each midpoint level impact (e.g.,

kg CO2 eq. of carbon dioxide for the Impact 2002+ global warming impact). The figures illustrate that the introduction of WPAs almost always decreases net impacts by augmenting avoided burdens and by reducing impacts. Of perhaps greatest note is the agreement of the three methods in identifying the reduction in climate change emissions as the most important contributor to the change in avoided burdens. Most of the midpoint results for equivalent categories tend to be quite similar (e.g., avoided burdens from ozone layer depletion: -9%). However, there are substantial differences between relatively similar impact categories such as the Impact 2002+ non- renewable energy category (including fossil fuels and uranium) and ReCiPe’s fossil fuel depletion (excluding uranium) category. The change in avoided burdens under ReCiPe is much greater than under Impact 2002+, mainly due to the exclusion of uranium in the former because energy from uranium was responsible for most of the avoided burden in the Impact 2002+ non-renewable energy category. The global warming/climate change impacts also produced a discrepancy primarily because Impact 2002+, unlike the other two selected methods, uses a 500 year time horizon to calculate global warming potential (Jolliet et al. 2003). 151

s ic cts n e cs gen n o y ni ry eff it e o c ication rcin t catio h a ifi ca ir d n sp ci A Global WarmingEcotoxi Eutrop Carcinog No Ozone depletioRe Smog 0%

-5%

-10%

-15%

-20%

-25%

-30%

-35%

-40%

-45%

-50% % Change in Impact % Change in Avoided Impact

Figure 26 Net percent change in the midpoint level environmental impacts from the 2008 reference scenario under the TRACI 2 LCIA method. A positive value represents an increase in the environmental impact and a negative value represents a decrease. The reference values for each impact category are available in Appendix 4. Note: the global warming avoided impact shows a decrease of 225%.

ity ics on tri c ergy n cs ti ity xi n ni /nu ic to on ion a ica o ens on le e g f tox g ati ati inorga or o di ab tract y l acid ec up w x ia arming ec ial ens c ory tr w og g ra l e at ator s tic n a ir i -carcino in oc pir uatic acidirre rc n z n-rene s sp q a ni nd o iner e e A Te Global Aqua TerrestrAquatic CeutrophicationNo Io La N M Ozone layerR depletionR 20%

-20%

-40%

-60%

-80%

-100%

-120%

-140%

-160% % Change in Impact % Change in Avoided Impact

Figure 27 Net percent change in the midpoint level environmental impacts from the 2008 reference scenario under the Impact 2002+ LCIA method. Note: the global warming avoided impact shows a decrease of 164%. 152

n io n n n mat io io io r n ity y n upat mat n rmat io it c r tio ant fo at io a d ic ity xic icat n nsfo up otoxicic o trophication io d oc a c n tter fol oxi idif c ot u y t n c on io a c nge c it ia ti t a d tr o le er e er e xic n d p letion pletion mica ial a ch t ial e t o rad n e p eple e e tr ecotoxtr eutroph t g ltural la la d d te e e n u al la r d ulate mch in r sil e ic to rres rres ma is s t o e e ban o zone T ClimaFreshwaMarin T FreshwaMarin Hu Ion Agric Natu Ur F Metal Wade O Part Ph 10%

-10%

-20%

-30%

-40%

-50% % Change in Impact % Change in Avoided Impact

Figure 28 Net percent change in the midpoint level environmental impacts from the 2008 reference scenario under the ReCiPe (H) LCIA method. Note: the “climate change” avoided impact shows a decrease of 220%.

Those processes that were the primary contributors to both increases and decreases in the midpoint impact category indicators are listed in Appendix 4. The landfilling of waste was the most important process that caused environmental burdens for the largest number of midpoint impact categories (climate change, freshwater eutrophication, marine eutrophication, human toxicity, freshwater ecotoxicity and marine ecotoxicity). In contrast, the gross avoided impacts from newsprint prevention and recycling were responsible for the greatest reductions in impacts for the highest number of midpoint impact categories. Unlike ReCiPe’s other midpoint level impact categories, the water depletion impact is not incorporated into any of the endpoint level categories. Under both scenarios, the water depletion impact estimates depict avoided potential water consumption of 8.3*105 m³ (2008 reference scenario) and 1.2*106 m³ (waste prevention scenario). In areas of water scarcity, the avoided water consumption benefits would be augmented. However, the local benefits from the net avoided consumption would depend 153

on the specific locations of the recycling, raw material, waste processing and product manufacturing facilities within the life cycle of the system studied.

5.6.2 Endpoint level comparisons

Although much can be assessed from midpoint results, they do not reveal the significance of each type of impact on the environmental burdens of each scenario. The endpoint indicators are necessary for such an analysis. ReCiPe’s endpoint indicators are: (1) damage to human health, expressed as human health disability-adjusted life year (DALY); (2) damage to ecosystems, measured as ecosystem species lost per year; and (3) reduction in resource availability, represented as the surplus cost of resources (Goedkoop et al. 2009). The Impact 2002+ LCIA method has roughly comparable endpoint indicators: (1) damage to human health (DALY), (2) damage to ecosystem quality (Potentially Disappeared Fraction of species – PDF*m2*yr), and (3) reduction in resource availability (MJ primary energy) (Jolliet et al. 2003). Impact 2002+ also considers climate change to be a separate endpoint indicator, while ReCiPe accounts for climate change within its set of three indicators. When evaluating the same LCA scenario using various LCIA methods, differences in the endpoint impact assessment results may be attributable to the following: (1) midpoint impacts are characterised differently within their respective endpoint impact categories; and (2) the endpoint impact categories and means of accounting for the impacts are not consistent. The net endpoint level impacts of the two scenarios are displayed in Table 16. When examining this table, it is immediately obvious that the results for similar indicators from the two LCIA methods can be quite different. This is especially the case for human health degradation and natural resource depletion. Although the impact results in Table 16 are grouped according to the type of impact indicator, these indicators are defined differently in the two LCIA methods, which has effects on the relative burdens of each of the LCA scenarios. For each endpoint level indicator, both LCIA methods reveal that the waste prevention scenario produces very substantial environmental and human health benefits relative to the 2008 reference.

154

Table 16 Net endpoint level impacts from the 2008 reference and waste prevention scenarios

Endpoint level Impact 2002+ ReCiPe (H) impact indicator 2008 reference net Waste prevention 2008 reference net Waste prevention impact1 scenario net impact impact1 scenario net impact 7 6 Climate change 2.41*10 kg CO2 eq. -5.14*10 kg CO2 eq. N/A N/A Ecosystem quality -5.26*107 -5.64*107 -3.43 species*yr -3.69 species*yr PDF*m2*yr PDF*m2*yr Human health -2.07 DALY -2.83*101 DALY 6.03*102 DALY 5.26*102 DALY Natural resources -9.97*108 MJ -1.55*109 MJ $1.20*108 -$4.07*107 1 PDF = Potentially disappeared fraction of species; DALY = Disability Adjusted Life Years

5.6.2.1 Damage to ecosystem quality

Although the ecosystem quality indicators are not strictly comparable because Impact 2002+ excludes climate change impacts from its indicator, the results using both LCIA methods show significant ecosystem quality benefits under both scenarios (Figure 29). The results also indicate that the shift from the 2008 reference to the waste prevention scenario produces net avoided impacts for ecosystem quality of 6.8% to 7.6%. Even in the waste prevention scenario, recycling is responsible for the vast majority of the ecosystem benefits. Although generating net environmental gains, the prevention of some waste material inputs to the recycling system diminishes the gains from the WPAs by 50.0% (Impact 2002+) to 50.9% (ReCiPe).

155

4.0 Impact 2002+ 4.0E+07 ReCiPe

2.0 2.0E+07

WPA 1 0.0E+00 0.0 WPA 2 WPA 3 -2.0E+07 WPA 4+5 -2.0 Recycling

Species*yr Landfilling

PDF*m²*yr -4.0E+07 Biological Treatment -4.0 Waste Transport -6.0E+07 Waste Collection

-6.0 -8.0E+07

-1.0E+08 -8.0 2008 Waste 2008 Waste Reference Prevention Reference Prevention Scenario Scenario

Figure 29 Damage to ecosystem quality endpoint level damage indicator values for the life cycle components of the 2008 reference and waste prevention scenarios. For all impact categories, the WPA and recycling components are responsible for the avoided impacts, while the waste collection, waste transport, biological treatment and landfilling components are responsible for the environmental damage.

Land occupation was responsible for all of the ecosystem quality gains (Figures 30 and 31). The acidification, ecotoxicity and eutrophication impacts were of relative insignificance. Although the land occupation midpoint level impact tends to have greater uncertainty than the other types of impacts (Humbert et al. 2005), it nevertheless has an overwhelming influence using both LCIA methods. ReCiPe revealed much more substantial impacts from landfilling than did Impact 2002+, due to the incorporation of climate change impacts on ecosystem quality.

156

2008 Reference

Land occupation Terrestrial acid/nutri Terrestrial ecotoxicity Aquatic ecotoxicity

Waste Prevention Scenario

0% 20% 40% 60% 80% 100% % of the sum of the absolute values of the Damage to Ecosystem Quality endpoint impacts

Figure 30 Percentage contributions of midpoint level impacts to the damage to ecosystem quality endpoint level damage indicator values, for the 2008 reference and waste prevention scenarios (Impact 2002+). Land occupation, unlike the other midpoint level impacts, is an avoided impact.

2008 Reference

Agricultural land occupation Urban land occupation Natural land transformation Climate change Freshwater ecotoxicity Terrestrial ecotoxicity Terrestrial acidification Waste Prevention Scenario

0% 20% 40% 60% 80% 100% % of the sum of the absolute values of the Damage to Ecosystems endpoint impact

Figure 31 Percentage contributions of midpoint level impacts to the damage to ecosystem quality endpoint level damage indicator values, for the 2008 reference and waste prevention scenarios (ReCiPe). Patterned and white-coloured bars represent avoided impacts. 157

5.6.2.2 Damage to human health

While using an identical unit of measurement (DALY), the endpoint characterisation factors to convert midpoint impacts to endpoint impacts differ significantly between Impact 2002+ and ReCiPe (H). The result using Impact 2002+ showed a net benefit while the ReCiPe result showed damages at an order of magnitude of one to two times that of Impact 2002+. The overall magnitudes of the DALYs primarily differed due to the exclusion of climate change from the Impact 2002+ damage to human health indicator, and the different methods used to calculate human toxicity, (including the human toxicity, carcinogen and non-carcinogen impact categories). The LCIA results tend to be consistent (with the exception of recycling) in terms of the relative importance of each of the human health impact sub-categories (Figure 32). Overall, the LCIA methods indicate considerable relative reductions in net human health damage due to the shift from the 2008 reference to the waste prevention scenario (i.e., 26 and 77 disability adjusted life years - DALYs - under Impact 2002+ and ReCiPe, respectively). Excluding climate change impacts, Impact 2002+ identified the greatest reductions in non-carcinogens, respiratory inorganics and carcinogens, while ReCiPe identified human toxicity and particulate matter formation. These impact reductions can be attributed in part to the decrease in upstream product/material transport associated with the WPAs. Although high uncertainty is associated with this impact indicator (Humbert et al. 2005) and the overall magnitudes of the impact estimates, the relative improvements in the waste prevention scenario using both RcCiPe and Impact 2002+ are comparable. For example, the WPAs are responsible for a 9% (ReCiPe) and 12% (Impact 2002+) reduction in absolute human health impacts (i.e., a sum of the absolute values of those impacts and avoided impacts from each component of the life cycle). 158

Impact 2002+ ReCiPe 8.0E+02 8.0E+02

7.0E+02 7.0E+02

6.0E+02 6.0E+02

5.0E+02 5.0E+02 WPA 1 4.0E+02 4.0E+02 WPA 2 WPA 3 3.0E+02 3.0E+02 WPA 4+5 DALYs DALYs Recycling 2.0E+02 2.0E+02 Landfilling Biological Treatment 1.0E+02 1.0E+02 Waste Transport Waste Collection 0.0E+00 0.0E+00

-1.0E+02 -1.0E+02

-2.0E+02 -2.0E+02 2008 Waste 2008 Waste Reference Prevention Reference Prevention Scenario Scenario

Figure 32 Damage to human health endpoint level damage indicator values for the life cycle components of the 2008 reference and waste prevention scenarios. For all impact categories, the WPA and recycling components are responsible for the avoided impacts, while the waste collection, waste transport, biological treatment and landfilling components are responsible for the environmental damage.

Using Impact 2002+, the relative contributions of the six midpoint impacts to the endpoint indicator, in absolute terms, are substantial only for respiratory inorganics, carcinogens, and non-carcinogens (percent contributions range between 23-49%). The results from ReCiPe, excluding climate change, which is responsible for approximately 70% of the absolute impact of the two scenarios, are in agreement with Impact 2002+ inasmuch as particulate matter formation (analogous to respiratory inorganics and respiratory organics) is a significant contributor to human health damage. In absolute terms, ReCiPe’s human toxicity indicator, and the carcinogens and non-carcinogens endpoint categories of Impact 2002+, also make considerable contributions. However, the Impact 2002+ indicators are responsible for avoided burdens while the ReCiPe indicator causes human health damage (Figures 33 and 34), an illustration of the high uncertainties associated with toxicological impacts (Winkler and Bilitewski 2007). 159

Moberg et al. (2005) exemplifies another LCA of waste which attempted to illustrate the uncertainty of toxicological impacts through comparisons using different impact characterisation methods

2008 Reference

Carcinogens Non-carcinogens Ionizing radiation Respiratory inorganics Respiratory organics Ozone layer depletion

Waste Prevention Scenario

0% 20% 40% 60% 80% 100% % of the sum of the absolute values of the Damage to Human Health endpoint impact

Figure 33 Percentage contributions of midpoint level impacts to the damage to human health endpoint level damage indicator values, for the 2008 reference and waste prevention scenarios (Impact 2002+). Patterned and white bars represent avoided impacts.

160

2008 Reference

Ionising radiation Climate change Human toxicity Particulate matter formation

Waste Prevention Scenario

0% 20% 40% 60% 80% 100% % of the sum of the absolute values of the Damage to Human Health endpoint impact

Figure 34 Percentage contributions of midpoint level impacts to the damage to human health endpoint level damage indicator values, for the 2008 reference and waste prevention scenarios (ReCiPe). Ionizing radiation, unlike the other midpoint level impacts, is an avoided impact.

5.6.2.3 Depletion of natural resources

The depletion of natural resource impacts, which incorporate both metals/minerals and fossil fuel/non-renewable energy depletion, are negative (i.e., net increase in supply) under both scenarios, with the exception of ReCiPe’s 2008 reference scenario result (Figure 35). Although using different units to depict impacts on resource availability, the non-renewable energy gains under both LCIA methods are approximately two orders of magnitude greater than the gains in metals/minerals (Figures 36 and 37). The introduction of the WPAs resulted in the fossil fuel depletion impact falling by 133% (from depletion to avoided fossil fuel use due to the fall in consumption) under ReCiPe. Similarly, although to a lesser extent, the avoided non-renewable energy consumption increased by 55% under Impact 2002+ (see Section 5.6.1 for explanation of discrepancy). For mineral supplies, results using Impact 2002+ showed a relative gain of 19%, with ReCiPe indicating 27%. 161

3.0E+09 6.0E+08 Impact 2002+ ReCiPe

2.0E+09 4.0E+08

WPA 1 WPA 2 1.0E+09 2.0E+08 WPA 3 WPA 4+5 Recycling

0.0E+00 $ 0.0E+00 N Landfilling Biological Treatment MJ primary energy Waste Transport -1.0E+09 -2.0E+08 Waste Collection

-2.0E+09 -4.0E+08

-3.0E+09 -6.0E+08 2008 Waste 2008 Waste Reference Prevention Reference Prevention Scenario Scenario Figure 35 Resource availability endpoint level damage indicator values for the life cycle components of the 2008 reference and waste prevention scenarios. For all impact categories, the WPA and recycling components are responsible for the avoided impacts, while the waste collection, waste transport, biological treatment and landfilling components are responsible for the environmental damage.

2008 Reference

Mineral extraction Non-renewable energy

Waste Prevention Scenario

0% 20% 40% 60% 80% 100% % of the sum of the absolute values of the Reduction in Resource Availability endpoint impact

Figure 36 Percentage contributions of midpoint level impacts to the Resource Availability endpoint level damage indicator values, for the 2008 reference and waste prevention scenarios (Impact 2002+). Both bars depict a net increase in resource supplies.

162

2008 Reference

Metal depletion Fossil depletion

Waste Prevention Scenario

0% 20% 40% 60% 80% 100% % of the sum of the absolute values of the Natural Resource Depletion endpoint impact

Figure 37 Percentage contributions of midpoint level impacts to the Resource Availability endpoint level damage indicator values, for the 2008 reference and waste prevention scenarios (ReCiPe). The patterned metal depletion bar represents an avoided impact while the fossil depletion bar with the shading gradient represents a net loss of the resource under the 2008 reference scenario, and a net gain in supply under the waste prevention scenario.

5.6.2.4 WPA comparison

All of the examples of the WPA types generated net avoided burdens, which indicates that the substitute product services (if necessary) are always responsible for lower net impacts than the product services they are replacing. The WPAs have different upstream and downstream effects within the waste prevention scenario. Table 17 and 18 list these endpoint level impacts and reveal that the downstream effects of the WPAs, except grasscycling, increase when one accounts for the lost environmental gains from recycling. In all cases, the life cycle benefits of the WPAs surpass the life cycle burdens. For the damage to human health indicator, the recycling losses displace 17% (Impact 2002+) of net avoided burdens from the WPAs, although for ReCiPe the benefits from recycling increase slightly (2%). For resource availability, 23% (Impact 2002+) and 3% (ReCiPe) of the net avoided burdens from the WPAs are displaced, while the displacement for Impact 2002+’s climate change indicator is 4%. The effect of recycling losses is most pronounced for the damage to ecosystem quality indicator, with a 59% 163

(Impact 2002+) and 66% (ReCiPe) displacement of the net avoided burdens from the WPAs. The larger effect on ecosystem quality is due to the high proportional importance of recycling to the results for this indicator.

Table 17 Net endpoint level impacts of each WPA within the waste prevention scenario (Impact 2002+)

Endpoint Upstream effect of WPA Downstream effect level impact WPA 1 WPA 2 WPA 3 WPAs 4 WPA 7 Downstream Recycling indicator (Admail) (Disposable (Newspapers) and 5 (Grasscycling) excluding bags) (Wine and recycling spirit packaging) Climate -9.67*106 -2.49*106 -3.28*106 -1.31*107 N/A -2.04*106 1.34*106 change (kg CO2 eq.) Ecosystem -4.15*106 -1.52*105 -1.65*106 -2.14*106 N/A -4.93*105 4.78*106 quality (PDF*m²*yr) Human -9.99 -2.08 -3.59 -13.0 N/A -2.54 5.00 health (DALY) Natural -3.17*108 -9.33*107 -1.02*108 -1.65*108 N/A -3.14*107 1.59*108 resources (MJ)

Table 18 Net endpoint level impacts of each WPA within the waste prevention scenario (ReCiPe (H))

Endpoint Upstream effect Downstream effect level impact WPA 1 WPA 2 WPA 3 WPAs 4 WPA 7 Downstream Recycling indicator (Admail) (Disposable (Newspapers) and 5 (Grasscycling) excluding bags) (Wine and recycling spirit packaging) Ecosystem -3.07*10-1 -2.43*10-2 -1.19*10-1 -1.39*10-1 N/A -5.86*10-2 3.89*10-1 quality (species*yr) Human -2.34*101 -4.75 -6.62 -2.80*101 N/A -1.26*101 -1.55 health (DALY) Natural -4.92*107 -3.19*107 -1.75*107 -5.55*107 N/A -1.15*107 5.18*106 resources ($)

5.6.2.5 Comparisons of absolute endpoint impact values

Figures 38 and 39 provide an illustration of the importance of each life cycle component of the 2008 reference and waste prevention scenarios as a percent of the sum of the absolute impact values, for each endpoint indicator. Absolute values are used for 164 these comparisons because there are impacts causing damage as well as avoided burdens in both LCA scenarios. The results for the 2008 reference scenario displayed in Figure 38 reveal the importance of recycling (13%-76% of absolute impacts, with 13% as an outlier) and MSW collection and transportation (7%-46% of impacts, with 7% as an outlier) within the life cycle of waste. The results also challenge the claim that the environmental burdens from waste transportation are negligible in comparison with waste treatment (e.g., Mendes et al. 2004). For all of the endpoint level indicators using Impact 2002+ and one of the indicators using ReCiPe, waste collection and transportation have considerably larger impacts than landfilling. Recycling and waste transportation from the transfer stations to the location of treatment are the most important components according to Impact 2002+, whereas ReCiPe identifies recycling, landfilling and waste transportation. The biological treatment system provides the lowest contribution of impacts under all impact categories but one (i.e., ecosystem quality under ReCiPe). Recycling, unlike the other components of the 2008 reference scenario life cycle, is responsible for a net impact benefit. In the 2008 reference scenario, the largest discrepancy between the results from the Impact 2002+ and ReCiPe methods is in their evaluations of the importance attributed to landfilling, with the former method indicating that it is always responsible for less than 11% of the impact, and the latter method generating results which vary from approximately 8% to 63%. This discrepancy can be attributed, in part, to the different time horizons used in the calculation of global warming potential and other impacts. Methane emissions from landfills are far more potent over a 100 year time horizon (RcCiPe) than under a 500 year time horizon (Impact 2002+). 165

ReCiPe

Resources

Ecosystem quality

Human health

Recycling Waste Collection Impact 2002+ Waste Transport Biological Treatment Landfilling Climate change

Resources

Ecosystem quality

Human health

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% % of the sum of the absolute values of the endpoint level damage indicators

Figure 38 Life cycle components of the 2008 reference scenario as a percent of the sum of the absolute values for the endpoint level damage indicators.

If one would attempt to depict the waste prevention scenario using the traditional system boundaries of an LCA of MSW, the impact contribution percentages of each life cycle component would show only a slight change from the 2008 reference scenario (Figure 38). However, the overall magnitudes of some of the impacts would be shown to increase with the incorporation of the WPAs. This apparently odd finding is caused by the effect of preventing the generation of mostly recyclable materials. Specifically, an asymmetry in the system boundary of the LCA would be produced because the avoided burdens from recycling would be taken into account, while the upstream avoided burdens of waste prevention would not. Thus, if one does not employ expanded boundaries when evaluating the effects of WPAs, the evaluation will be misleading. Figure 39 depicts the waste prevention scenario, which incorporates the upstream effects of waste prevention into the results through the expansion of the LCA system boundaries relative to the traditional LCA of MSW. The waste prevention scenario 166 results indicate that, for every endpoint indicator, the downstream components have a greater contribution than the net changes upstream. This is not surprising, since the quantity of waste prevention undertaken in the scenario is equivalent to only 3.5% of the amount of waste collected by the City of Toronto in 2008. However, the net upstream benefits of these WPAs are always proportionally greater than the percent mass reduction would suggest (Note: the grasscycling WPA has only downstream benefits). For five of the seven endpoint level indicators, these upstream benefits displace impacts greater than those resulting from the landfilling and biological treatment of all of the City of Toronto’s residential waste. Among the various WPAs included in the waste prevention scenario, the net upstream benefits from the prevention of admail and conventional wine and spirit containers provide the greatest contribution to reducing the impacts of residential waste management in Toronto.

ReCiPe

Resource availability

Ecosystem quality WPA1

Human health WPA2

WPA3

Impact 2002+ WPA4 and 5

Recycling Climate change Waste Collection

Resources Waste Transport

Biological Treatment Ecosystem quality Landfilling

Human health

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% % of the sum of the absolute values of the endpoint level damage indicators

Figure 39 Life cycle components of the waste prevention scenario as a percent of the sum of the absolute values for the endpoint level damage indicators. For all impact categories, the WPA and recycling components are responsible for the avoided burdens, while the waste collection, waste transport, biological treatment and landfilling components are responsible for the environmental damage.

167

5.7 Sensitivity analysis

The avoided burdens of recycling are evaluated using the assumption that virgin sources are displaced (see Section 5.5.2.4), while the avoid burdens from waste prevention result from the avoided production of materials from both recycled and virgin sources. These assumptions can lead to the perhaps implausible result which shows net increases in environmental effects when implementing some WPAs affecting the production of goods with recycled content. For example, excluding collection, sorting and processing losses, the prevention of a newspaper with 50% recycled content would reduce forest harvesting for the newsprint by 50%. In comparison, recycling the newspaper would eliminate forest harvesting for the inputs of a new newspaper, which would therefore produce a more favourable outcome in terms of land use. However, this example does not account for the likelihood that the recycled component of the prevented newspaper would have displaced the virgin component of another newspaper. Since the recycled content in the prevented materials is taken into account in this LCA, it is important to undertake a sensitivity analysis to determine the effect of assuming no recycled content in the materials subject to waste prevention in the waste prevention scenario. The results displayed in Table 19 illustrate the significance of the recycled content assumption on the upstream net avoided burdens of the WPAs. If it is assumed that the materials subject to waste prevention contain no recycled content, the increase in the net avoided impacts rise by up to 43%, with the greatest improvements in the ecosystem quality indicator due to the reduction in land exploitation to supply pulp.

Table 19 Percentage changes in the endpoint level net avoided impacts from the default WPAs to those in an altered scenario in which the materials subject to waste prevention contain no recycled content

Endpoint level Impact 2002+ ReCiPe (H) impact indicator “No recycled content” Percent increase in net “No recycled content” Percent increase in net waste prevention avoided impacts from waste prevention avoided impacts from scenario impact1 the upstream scenario impact1 the upstream component of the component of the WPAs WPAs 6 Climate change -6.72*10 kg CO2 eq. 5.6% N/A N/A Ecosystem quality -5.95*107 PDF*m2*yr 38.3% -3.95 species*yr 43.2% Human health -3.17*101 DALY 12.0% 5.26*102 DALY 0.4% Natural resources -1.68*109 MJ 19.0% -$5.02*107 6.2% 1 PDF = Potentially disappeared fraction of species; DALY = Disability Adjusted Life Years 168

Due to the significant contribution of waste transportation to the endpoint level life cycle impacts of the 2008 reference scenario, a sensitivity analysis is undertaken to substitute the Green Lane landfill (203 km from Toronto) for the Carleton Farms landfill in Michigan (441 km). As of 1 January 2011, all residential waste from the City of Toronto that previously was transported to the Michigan landfill is now being sent to the Green Lane landfill, which is 238 km closer to Toronto. Since the transportation component of the 2008 reference scenario is responsible for a substantial contribution to the life cycle impacts of MSW management (5%-32% - see Figure 38), it is of interest to evaluate the effect of this reduction in the average transportation distance of the waste. Table 20 displays the changes in the endpoint level impacts of waste transportation in 2008 reference scenario, substituting the Green Lane landfill for the Carleton Farms landfill in Michigan. The beneficial effects of this substitution are impressive using either LCIA method, and are roughly equivalent to half of the percentage changes for several endpoint impacts resulting from the WPAs. The results in the table indicate a net reduction in waste transportation impacts of approximately one third.

Table 20 Percentage changes in the endpoint level waste transportation impacts from the 2008 reference scenario to an altered scenario in which all landfilled waste is sent to the Green Lane landfill in St. Thomas, Ontario

Endpoint level impact indicator Impact 2002+ ReCiPe (H) % impact change from 2008 reference % impact change from 2008 reference for waste transportation for waste transportation Climate change -33.2% N/A Ecosystem quality -35.5% -33.1% Human health -29.7% -31.5% Natural resources -33.8% -33.9%

5.8 Critical review

The International Organization of Standardization (ISO) guidelines for LCA outline numerous criteria which must be met when undertaking a comparative study intended for public disclosure. The LCA must be scientifically and technically valid, as well as environmentally relevant inasmuch as there should be linkages to category endpoints (ISO 2006). This critical review addresses the interpretation phase of the LCA. 169

Uncertainties identified in this review are related to the accuracy and precision of the key input data, as well as with the LCIA method applied. Upstream data inputs for this LCA included the reference flows of product systems removed and added due to the WPAs, as well as the types of unit processes used to depict these systems. These LCA data inputs were based on measurements and assumptions that have uncertainty, although the uncertainty differs according to the intended scope and purpose of the LCA. For example, the results are far less uncertain if they are intended to depict the prevention of 104 tonnes of unaddressed advertising mail waste, rather than the prevention of admail for the 65% of residents (based on a published estimate from a survey with its own inherent uncertainty) who claimed that such advertising is of no interest to them. When the latter depiction is used, one must account for the additional uncertainty associated with the surveyed percentage of the Canadian population who found admail of no interest to them (FDSA 2007) and the admail opt-out estimate from Canada Post (2007). When the former depiction is adopted, the only uncertainty associated with the magnitudes of reference flows of product systems removed and added due to the WPAs pertains to the substitution levels of alternate product systems. For the newspaper WPA example, there is uncertainty associated with the 20 minute estimate of the average daily length of time that the average Toronto resident would spend reading newspaper articles. In many cases, it has not been possible to determine the levels of uncertainty because the sources for the input data did not examine this issue. Unlike the other data inputs into the LCA model, any uncertainty associated with the waste prevention tonnage estimates would not take into account differences between the estimates and existing circumstances. In all cases, the waste prevention estimates were calculated as percentages of particular waste streams which would be subject to a WPA. The plausibility of each of these waste prevention estimates was addressed in detail in Sections 5.3.1 and 5.3.4, as well as in Appendix 4. Although for calculation purposes in the model the waste prevention estimates were rounded to the tonne, it was stated in footnotes to Tables 12 and 14 that the specific number of digits associated with these estimates were not intended to indicate a particular level of certainty in terms of the magnitudes of existing WPAs. 170

The sources of downstream data inputs included the City of Toronto, Stewardship Ontario, and Waste Diversion Ontario. Data with the lowest levels of uncertainty, perhaps only a few percent (i.e., due to the possibility of an inaccurate weighing scale or a missed truck), were the quantities of waste generated, which were based on weighed truckloads. Although no uncertainty data is available for waste tonnage measurements, it is assumed that such measurement errors were relatively small due to the need for accurate calculations of landfill tipping fees, as well as prices for recyclable materials and organic waste. These revenues/costs are often based on the number of tonnes supplied/purchased. Published waste audits based on 100 to 200 samples of household waste (per season, by type of dwelling) in Toronto were used to depict the composition of landfilled waste. Although the tonnage landfilled has low uncertainty, the uncertainty associated with waste composition for specific waste streams such as daily and weekly newspapers, would be higher. There was insufficient information provided to statistically analyze the uncertainty of the estimates for specific waste streams. Nevertheless, the potential error associated with waste composition was limited by the measured tonnage of all waste landfilled (i.e., if the tonnage of one waste stream is underestimated, it must be counterbalanced by a reduction in the tonnage estimates of the remaining waste streams). Waste transportation distances and tonne-km would have relatively low uncertainty because the majority of the collected waste is known to be sent by the municipality to particular landfills, anaerobic digestion and composting facilities. However, the transportation distances for recyclable materials would have somewhat higher levels of uncertainty due to less certain data collected by brokers who purchased recyclable materials from several sources and sold them to numerous customers (i.e., the original source of the final market data has an additional degree of separation: the data was provided by the broker to the municipality). Nevertheless, the tonnage of recyclable waste was significantly less than the tonnage of landfilled and biologically treated waste, which would reduce the effect of this uncertainty on the overall results. Some system boundaries were inconsistent in the accounting of capital goods. Those environmental burdens from the production of capital goods, including infrastructure, were included in most unit processes. However, some capital goods were 171 excluded due to a lack of available data. Examples of exclusions included machinery and facilities for bottle washing and filling, aseptic carton manufacturing, PET and HDPE recycling and organic waste treatment. It seems likely that effects from the omission of capital goods in these few unit processes would have produced a negligible contribution to the total impacts of the scenarios. This assertion is made in light of the observations of Frischknecht et al. (2007) that the environmental burdens from capital goods often tend to be insignificant. Uncertainty was associated with the choice of unit processes entered into the LCA model to represent the waste management system. There were a number of unit processes in which the geography of the technology was not identical to that in which the process occurred. The substitution of more representative electricity mixes into these unit processes was undertaken to compensate for such geographical discrepancies. The magnitudes of some impacts depended on the location of the environmental emission (or in the case of resource availability, the location of resource consumption). For example, the impact of water use in an area of water scarcity would be greater than in a location of water abundance. The climate change and resource availability impact categories are relatively insensitive to the geographic location of the environmental intervention. In contrast, impacts from air, water and land pollutants, such as human toxicity, ecotoxicity, acidification, and eutrophication, are sensitive. However, the regionalization of impacts remains in its infancy, with no regionalized impact assessment factors used in LCA databases or large LCA software packages (Mutel and Hellweg 2009). The age of the data used in the LCA was not always consistent or representative of the conditions in 2008. Some unit processes, such as the production of computers for the newspaper WPA, may have been out of date due to their fast pace of change. These types of issues are present in most LCAs, and are addressed through methodological transparency, and, where possible, through various mitigation measures. For example, the LCA processes defined in the US-EI database use, by default, the US electricity mix from 2004. For this LCA, these US-EI unit processes were replaced with a more appropriate electricity mix reflecting 2008 conditions (see Appendix 4). 172

There was considerable uncertainty inherent to the landfilling unit process. This uncertainty was associated with the composition of the landfilled waste and the modelling of the short-term (less than 100 years) and long-term (up to 60 000 years) transfer coefficients of chemical elements (Doka 2009). Due to the substantial contribution of the landfilling component of the life cycle in terms of its absolute endpoint level impacts (i.e., from less than 10% of impacts under the Impact 2002+ method to over 60% of human health impacts under ReCiPe), it is prudent to identify and evaluate the uncertainties of the landfilling unit process. In the documentation for the landfilling unit process for EcoInvent, Doka (2009) contends that the composition of waste inputs to landfill tends to be more uncertain than the modelled chemical element transfer coefficients. The uncertainty of waste inputs is especially evident for those trace elements, such as copper and cadmium, which can be responsible for considerable impacts. At a confidence level of 95% (geometric standard deviation, GSD²) for the processes at a sanitary landfill, energy and land use estimates had an uncertainty of up to 150%, with processes related to infrastructure reaching 300% (Doka 2009). Moreover, leachate volume uncertainty (GSD²) in the landfilling unit process was 105% (Doka 2009). However, these impact estimates are small relative to those from the decomposition of landfilled waste and the transfer of chemicals via leachate and air emissions. Doka (2009) suggests that the levels of uncertainty associated with the transport of chemicals with large transfer coefficients (e.g., water soluble elements that are rapidly transferred to leachate in substantial quantities) are less than the levels for the transport of chemicals with low coefficients. For the chemical elements with high coefficients, the GSD² can reach 100% (Doka 2009). For those with low coefficients, often trace elements, the GSD² reaches as high at 1000% (Doka 2009). The high uncertainties of the potential impacts of landfilling reflect the inherent variability of landfilling processes, which are affected by waste composition, landfill design and operation, climate, and topography. Unlike the human health and ecosystem quality endpoint level impact estimates using Impact 2002+, those estimates from ReCiPe were mostly comprised of climate change impacts, which have less uncertainty than human 173

toxicity impacts. This disparity is a consequence of the Impact 2002+ method isolating climate change impacts from the other endpoint level impact categories. Although there was insufficient information provided in the landfill unit process documentation to estimate uncertainty, one can explore the potential effects of this uncertainty on the life cycle impacts of each component of the life cycle. For example, if the impact of landfilling would be much higher than the original estimate, this change would have no effect on the impacts from waste transportation. Therefore, the relative importance of the transportation life cycle component would diminish. In contrast, the impact benefits from WPAs could increase in relative importance due to the reduction in landfilled waste. For the newspaper WPA, significant uncertainty could be associated with the impacts from the waste management of the 10 additional tonnes of computer equipment used for reading newspaper articles online. Although modern waste management techniques were assumed, a substantial portion of this waste may have been landfilled or shipped to other countries which may not use such techniques (Carter-Whitney and Webb 2008), augmenting environmental and human health impacts. The additional amount of computer waste generated was quite small relative to the total amount of MSW addressed in this LCA, as well as to the tonnage of waste prevented in the newspaper WPA. Nevertheless, the application of inappropriate waste management techniques for this electronic waste would decrease the net life cycle benefit from implementing this particular WPA, especially given the toxicity of substances contained in electronic waste. Uncertainty increases significantly when interpreting results from endpoint level impacts instead of midpoint level ones (Goedkoop et al. 2009). Some LCIA methods such as ReCiPe group their assumptions and uncertainties into perspectives (i.e., egalitarian, individualist, and hierarchist) (Goedkoop et al. 2009). The documentation of the Impact 2002+ LCIA method (Humbert et al. 2005) identifies the relative levels of uncertainty associated with the midpoint and endpoint level indicators through a qualitative comparison. Midpoint level effects with low uncertainty include global warming potential, aquatic acidification and eutrophication, as well as non-renewable energy consumption. High uncertainties are associated with the impact categories of human toxicity, respiratory inorganics, ionizing radiation, aquatic and terrestrial 174 ecotoxicity, terrestrial acidification/nutrification, and land occupation effects. Ozone layer depletion and mineral extraction are considered impacts with medium uncertainty. Uncertainty associated with the resource availability endpoint level impacts is low, while that for the human health and ecosystem quality impact is estimated to be high. For the LCA endpoint level results for the 2008 reference and waste prevention scenarios, land occupation, which has high uncertainty according to Humbert et al. (2005), contributed most to the indicators of “damage to ecosystem quality.” ReCiPe‘s indicator of damage to human health was affected mainly by climate change, a low uncertainty impact, whereas the indicator for Impact 2002+ was comprised mainly of impacts with high uncertainty, such as respiratory inorganics and carcinogens (human toxicity). The resource availability indicators were almost entirely comprised of the non- renewable energy/fossil fuel depletion impacts, which have low uncertainty. The production of a quantitative estimate of the overall uncertainty of LCA results is often not feasible due to the magnitude of data inputs in LCA and LCIA models, each input having its own uncertainty and potential effect on the results (Franklin Associates 2010). Equally important for this LCA was the necessary use of certain single point estimates (e.g., waste tonnage) because they were the best or only data available, even though they may have lacked uncertainty information. Nevertheless, a number of the uncertainties associated with the key data inputs, unit process design, and LCIA impact categories have been examined in this review. In other cases, data inputs were contrasted with published estimates used in comparable studies (e.g., comparing waste prevention estimates for advertising mail). Those components of the LCA with higher uncertainties were addressed in other sections of this chapter through sensitivity analyses and by comparing results using multiple LCIA methods, which is in keeping with published LCAs of waste management systems that possess relatively robust uncertainty analyses, such as Moberg et al. (2005). Finnveden et al. (2005) claim that the uncertainties associated with the choice of technology and LCA methodology, including the simplification of complex impact processes, tend to be larger than those from data inputs. In this LCA, the representation of the data inputs and LCA unit processes (excluding landfilling) tends to have relatively low uncertainty. Substantially higher uncertainties are associated with the emissions 175

from the landfilling unit process, the emissions from the treatment of the additional electronic waste from the newspaper WPA, and the calculation of potential impacts during the impact assessment stage of the LCA. The degree of the difference in the estimated midpoint level impacts for the two scenarios under each LCIA method were comparable (Section 5.6.1). The endpoint level impacts diverged to a greater extent, especially due to the role of different time horizons used to calculate global warming potential. Despite the uncertainties, under none of the endpoint level impact categories using Impact 2002+, ReCiPe and TRACI 2 did the 2008 reference scenario display a higher potential impact than the waste prevention scenario.

5.9 Discussion

Using the traditional system boundary of an LCA of MSW, only the downstream impacts of waste prevention on the environmental performance of the MSW management system are taken into account. As was shown in Section 5.7.2, this boundary limitation may lead to results indicating that the WPA causes an increase in impacts. One can observe such an increase in impacts because the system boundary would take into account the reduction in waste inputs for recycling, as well as the decrease in the avoided burdens that would have been associated with that recycling. More importantly, the absence of the upstream waste prevention effects from the LCA of MSW results in the exclusion of the majority of the benefits of waste prevention. The upstream processes associated with the WPAs must be incorporated into the LCA of MSW in order to have a more realistic evaluation of the impacts of waste prevention. In using the system boundary expansion outlined in the WasteMAP model, the results of this LCA have clearly demonstrated the impressive potential for waste prevention activities to reduce the impacts of residential waste management. However, this system boundary expansion can make undertaking this type of LCA of MSW very data intensive. For this case study, not only were data needed for the waste management life cycle from waste generation to disposal (including landfilling, biological treatment and recycling), but also for those product life cycles associated with the waste prevention activities. 176

The LCIA results were broadly consistent between ReCiPe (H), Impact 2002+ and TRACI 2, in terms of the potential net impacts of the waste management life cycles. However, result differences were evident and mainly attributable to the following: (1) some LCIA methods excluded significant emissions from the calculation of the impact category indicators; (2) during the characterization stage, the LCIA methods were not consistent in assigning the same weights inventory results within the midpoint impact categories, and midpoint impacts were similarly assigned different weights within endpoint categories; (3) the effects of temporal dynamics on the impacts from the process emissions were modelled differently (e.g., modelling the environmental persistence of pollutants); and (4) the impact categories and units selected to account for the impacts were not consistent. The initiatives evaluated in the waste prevention scenario are either currently being pursued to some extent (the examples of WPAs 1, 2, 3, 5 and 7), or are under serious discussion within waste management circles, such as wine bottle refilling (e.g., Leighton 2010). The results from this LCA, based on the WasteMAP model, demonstrate the potential of the various types of WPAs to improve the environmental performance of waste management systems without compromising the quantity of product services available for consumption by municipal residents. The LCA results can be used to help make informed policies and initiatives designed to encourage municipal residents and commercial enterprises to reduce the quantities of waste that they generate. Waste prevention through plastic bag reuse, grasscycling and online newspaper reading can be facilitated through education campaigns and financial incentives. Other means of waste prevention, such as advertising mail opting out schemes and bottle refilling programs, often require additional institutional support. Waste prevention can also be initiated at the product design stage, through lightweight packaging design, for example. The admail and newspaper WPAs, as well as the substitution of lighter weight and refillable containers for wines and spirits were the most effective means of reducing endpoint level impacts through waste prevention. Out of the WPAs evaluated within the waste prevention LCA scenario, the plastic bag WPA tended to have the smallest upstream effect on the waste management system. The grasscycling WPA had only 177

downstream effects which comprised a very small portion of the waste collection (<1% by tonnage), transportation (<1% reduction in tonne-km) and biological treatment (<5% reduction in organic waste inputs) components of the life cycle. The sensitivity analysis highlighted the significance of the recycled content of the prevented waste streams on the net benefits of waste prevention. If the materials subject to waste prevention in the waste prevention scenario did not have any recycled content, the net upstream endpoint level benefits from the WPAs would have increased by up to 43%. This result appears to indicate that there are additional benefits from targeting WPAs to products with low recycled content. However, such additional benefits are diminished if the prevented recycled material inputs are reallocated to displace the virgin content of the products that continue to be produced. Outside of waste prevention, the LCA scenario results indicate that focusing on improving the environmental performance of systems to collect and transport waste would be effective in reducing endpoint level impacts due to the relative importance of these life cycle components. Specifically, the sensitivity analysis results illustrated the substantial benefits from utilizing a landfill with a much closer proximity to Toronto than the one used for most of Toronto’s waste disposal in 2008. Possible means of mitigating impacts from waste collection and transportation include switching transportation modes from truck to train, improving the fuel efficiencies of the transportation technologies employed, and reducing transport distances to waste treatment facilities.

5.10 Conclusion

The results of this case study of the WasteMAP LCA are in keeping with the assumption inherent in the waste hierarchy that waste prevention generally has a superior environmental performance to waste treatment options. These promising results for waste prevention indicate that further research into the potential of WPAs to reduce the life cycle effects of residential waste management is merited.

178

5.11 References

Association of Canadian Distillers. 2008. Annual statistics.

Bare J, Norris G, Pennington D, McKone T. 2003. TRACI: The Tool for the Reduction and Assessment of Chemical and Other Environmental Impacts. J Ind Ecol 6(3-4): 49- 78.

Canada Post. 2007. Newsroom - Letters to the Editor (Published letter by François Legault, Manager, National Media Relations, Canada Post). Available at: http://www.canadapost.ca/cpo/mc/aboutus/news/letters/brampton_admail.jsf. Accessed on 2011 03 31.

Canada Post. n.d. Instant answers. How do I stop delivery of admail items? Available at: http://www.canadapost.ca/cpo/mc/personal/support/helpcentre/receiving/ admail_stop.jsf. Accessed on 2011 01 16.

Canadian Newspaper Association. 2010. Circulation data report 2009. Available at: http://www.newspaperscanada.ca/system/files/2009CirculationDataReport_3.pdf. Accessed on 2010 01 15.

Canadian Plastics Industry Association (CPIA). n.d. Plastic bags: Reuse & recycling. Available at: http://www.plastics.ca/Recycling/PlasticBags/ReuseRecycling/index.php. Accessed on 2011 01 16.

Canadian Vintner’s Association. 2008. Annual statistics.

Canwest News Service. 2008 (31 May). Money-making junk mail the wrong shade of green for Canada Post. Available at: http://www.canada.com/montrealgazette/news/story.html?id=fadfe47a-e528-46d7-bbbb- 654c890e57e7. Accessed on 2011 03 31.

Carter-Whitney M; Webb C. 2008. Waste bytes! Diverting waste electrical and electronic equipment in Ontario. Canadian Institute for Environmental Law and Policy. Available at: http://www.cielap.org/pdf/EwasteOntario.pdf. Accessed on 2011 10 28.

City of Toronto. 2011. 2011 Community Environment Days. Available at: http://www.toronto.ca/environment_days/. Accessed on 2011 10 31.

City of Toronto. 2010a. Personal communication. 2010 06 30 and 2010 12 23.

City of Toronto. 2010b. City of Toronto: Solid Waste Management - Green Bin Program. Frequently asked questions. Available at: http://www.toronto.ca/greenbin/faq.htm. Accessed on 2010 08 21.

179

City of Toronto. 2008. 2008 breakdown by material of residential waste diversion. Available at: http://www.toronto.ca/garbage/pdf/2008-chart.pdf. Accessed on 2010 08 21.

City of Toronto. 2007. Proposed Initiatives and Financing Model to Get to 70% Solid Waste Diversion by 2010. Prepared for the City of Toronto Executive Committee by City of Toronto Solid Waste Management Services. Reference No. p:/2007/swms/may/011EC.doc. Available at: http://www.toronto.ca/legdocs/mmis/2007/ ex/bgrd/backgroundfile-3799.pdf. Accessed on 2010 09 15.

City of Toronto. 2006. Backgrounder: Release of 2006 Census results: Population and Dwelling Counts. Available at: http://www.toronto.ca/demographics/pdf/ 2006_population_and_dwelling_count_backgrounder.pdf. Accessed on 2010 09 15.

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):579-589.

Cox, J; Giorgi, S; Sharp, V; Strange, K; Wilson, D; Blakey, N. 2010. Household waste prevention — a review of evidence. Waste Manag Res 28: 193-219.

Doka G. 2009. Life cycle inventories of waste treatment services. Ecoinvent report no. 13. Swiss Centre for Life Cycle Inventories. Dubendorf, Switzerland. 133 pp.

Ekvall T, Assefa G, Bjorklund A, Eriksson O, Finnveden G. 2007. What life-cycle assessment does and does not do in assessments of waste management. Waste Manage 27: 989-996.

European Commission. 2005. Impact assessment on the thematic strategy on the prevention and recycling of waste. COM 666 final.

European Commission Joint Research Centre (EC JRC). 2010. ILCD Handbook: Analysing of existing Environmental Impact Assessment methodologies for use in Life Cycle Assessment. First edition. Available at: http://lct.jrc.ec.europa.eu/pdf- directory/ILCD-Handbook-LCIA-Background-analysis-online-12March2010.pdf. Accessed on 2011 04 30.

Exchange Magazine. 2006. Plastic Shopping Bags Too Valuable To Be Going To Waste - Industry Wants Them Back. Available at: http://www.exchangemagazine.com/ XQuarterly/Archive-associations-27-Mar06.html. Accessed on 2011 01 27.

Flyer Distribution Standards Association (FDSA). 2007. Reach (monthly newsletter of the FDSA, in partnership with the Retail Council of Canada). Available at: http://www.fdsa-canada.org/newsletter/december_2007.pdf. Accessed on 2011 02 09.

180

Goedkoop M, Heijungs R, Huijbregts M, De Schryver A; Struijs J, van Zelm R. 2009. ReCiPe 2008: A life cycle impact assessment method which comprises harmonised category indicators at the midpoint and the endpoint level. First edition. Report I: Characterisation. Rumte en Milieu. Ministerie van Volkshuisvesting. Ruimtelijke Ordening en Milieubeheer. Available at: http://s3.amazonaws.com/ jef.mindtouch.com/10059895/11/0?AWSAccessKeyId=1TDEJCXAPFCDHW56MSG2 &Signature=/M2Jz8%2b3G0vMifS6QJM3bieLwRc%3d&Expires=1282315930. Accessed on 2010 08 20.

Hischier R and Reichart I. 2003. Multifunctional electronic media – traditional media: The problem of an adequate functional unit. Int J LCA 8(4): 201-208.

Hung M-L, Ma H-w. 2009. Quantifying system uncertainty of life cycle assessment based on Monte Carlo simulation. Int J LCA 14(1): 19-27.

Jolliet O, Margni M, Charles R, Humbert S, Payer J, Rebitzer G, Rosenbaum R. 2003. IMPACT 2002+: A new life cycle impact assessment methodology. Int J LCA 8(6): 324-330.

181

International Energy Agency (IEA). 2010. IEA Energy Statistics – For Electricity/Heat (2008). Available at: http://www.iea.org/stats/prodresult.asp?PRODUCT =Electricity/Heat. Accessed on 2010 12 27.

International Organization of Standardization (ISO). 2006 Environmental management - Life cycle assessment - Requirements and guidelines. ISO 14044: 2006.

Leighton, C. 2010. Refillable wine bottles. Solid Waste & Recycling. October/November 2010 issue: 8-16. Available at: http://www.solidwastemag.com/issues/de.aspx?id=5474. Accessed on 2011 02 05.

Liquor Control Board of Ontario (LCBO). 2008. 2008 Volume sales data for wine and spirits. Unpublished.

Moberg Å, Johansson M, Finnveden G, Jonsson A. 2010. Printed and tablet e-paper newspaper from an environmental perspective - A screening life cycle assessment. Environ Impact Asses 30:3:177-191.

Mutel C, Hellweg S. 2009. Regionalized Life Cycle Assessment: Computational Methodology and Application to Inventory Databases. Environ Sci Technol 43:5797- 5803.

Ontario Plastic Bag Reduction Task Group. 2008. First Annual Report. Available at: http://www.wdo.ca/files/domain4116/Annual%20Report%20-%20Plastic%20Bag% 20Task%20Group%20-%20Dec08.pdf. Accessed on 2011 01 16.

Rigamonti, L. Grosso, M., and Caterina, M. 2009. Influence of assumptions about selection and recycling efficiencies on the LCA of integrated waste management systems. Int J LCA 14: 411–419.

Salhofer S, Obersteiner G, Schneider F, Lebersorger S. 2008. Potentials for the prevention of municipal solid waste. Waste Manage 28: 245-259.

Scarborough Research and Newspaper National Network LP. 2010. Telling the whole story: Analysis Supports Readership as Key Metric for Planning and Buying Newspaper Advertising. Available at: http://www.scarborough.com/press_releases/Scarborough% 20NNN%20Study%20Design%20Draft%20FINALa.pdf. Accessed on 2011 02 02.

Statistics Canada. 2008a. CANSIM. Table 051-0046. Catalogue no. 91C0029.

Statistics Canada. 2008b. CHASS Trade Analyser: Canadian Imports Data. 21 Dec 2008. HS10 Codes Used for Imported Bottled Wines: 2204101000, 2204109000, 2204211010, 2204211091, 2204211092, 2204211099, 2204212100, 2204212200, 2204212310, 2204212390, 2204212410, 2204212420, 2204212490, 2204212510, 2204212520, 2204212590, 2204212610, 2204212620, 2204212630, 2204212690, 182

2204212710, 2204212720, 2204212790, 2204212800, 2204213110, 2204213120, 2204213130, 2204213190, 2204213210, 2204213220, 2204213230, 2204213290, 2205101010, 2205101020, 2205101030, 2205101090, 2205102000, 2205103000, 2206002130, 2206002200, 2206003100, 2206003900, 2206004100, 2206004900, 2206005011, 2206005019, 2206005090, 2206006100, 2206006200, 2206006300, 2206006410, 2206006490, 2206006500, 2206006600, 2206006700, 2206006800, 2206007100, 2206007200.

Toffoletto L, Bulle C, Godin J, Reid C, Deschênes L. 2007. LUCAS – A New LCIA Method Used for a CAnadian-Specific Context. Int J LCA 12(2):93-102.

Toronto Star. n.d. Frequently asked questions. Available at: http://www.thestar.com/faq#printing. Accessed on 2011 01 22.

183

Valiante U. 2007. A Look at Tetra Pak's New Life-Cycle Inventory. Solid Waste & Recycling. February/March 2007.

Verghese K, Lewis H, Fitzpatrick L, Hayes G, Hedditch B. 2009. Environmental impacts of shopping bags. Ref number: SPA1039WOW-01. Available at: http://mams.rmit.edu.au/r97dgq3iero9.pdf. Accessed on 2011 01 17.

Weber C, Koomey J, Matthews H. 2010. The Energy and Climate Change Implications of Different Music Delivery Methods. J Ind Ecol 14(5):754–769.

Waste Diversion Ontario (WDO). 2008. Municipal Datacall: Tonnage. Available at: http://www.wdo.ca/content/?path=page82+item35931. Accessed on 2010 05 29.

Winkler J, Bilitewski B. 2007. Comparative evaluation of life cycle assessment models for solid waste management. Waste Manage 27:1021–31.

184

CONCLUSIONS

185

6.1 Summary of research findings

The four papers of this dissertation explore themes related to waste prevention, the system boundaries, functional units and scale of life cycle assessments (LCAs) of municipal solid waste (MSW) management, as well as the transparency and consistency of the application of LCA methods. They also examine the current methods and assumptions used in a large sample of published LCAs which evaluated the environmental performance of MSW management systems; propose a conceptual model to incorporate waste prevention activities into LCAs of MSW; compare product LCA results at individual product and municipal scales; and produce an LCA using the WasteMAP LCA conceptual model. This dissertation is able to provide a substantive answer to the primary research question which it posed: how can one evaluate the environmental burdens of a waste management system that incorporates waste prevention activities (WPAs)? It also meets the research objectives outlined in Chapter 1: (1) to quantify and compare the current methodological assumptions and results of LCAs of municipal solid waste management systems recently (2002-2008) published in peer-reviewed journals; (2) to propose a conceptual LCA model for evaluating and comparing, on a functionally equivalent basis, waste management systems that address different quantities of waste; (3) to produce LCAs of packaging systems (responsible for varying levels of waste generation) at the scale of the individual package and at the municipal scale, and to identify the ways in which the results of these LCAs differ; and (4) to evaluate and compare the results of a traditional LCA of MSW and an LCA scenario incorporating waste prevention activities, using the conceptual Waste Management And Prevention (WasteMAP) LCA model.

6.1.1 Life cycle assessments of municipal solid waste management systems: A comparative analysis of selected peer-reviewed literature (Paper 1)

The first paper addresses the first research objective of the dissertation. Its comparative analysis examines the methodological assumptions and results of published LCAs of waste management systems, and reveals the extent of the inconsistencies and 186 lack of transparency of the LCAs which are reviewed. In terms of the number of reviewed LCAs, Paper 1 remains the largest comparative analysis of LCAs of mixed- waste MSW management systems published in an English language peer-reviewed journal. This comparative analysis indicates the considerable extent to which the system boundaries of the reviewed LCAs differ. Some (3/20) exclude environmental burdens from waste transport. More than half of the LCAs either do not mention or are unclear if life cycle emissions from capital equipment are included in the calculation of results. Critical reviews of the published LCA results by means of explicit sensitivity analyses are not common, as they are present in only 4 of the 20 LCAs examined. Although economic impact indicators, via life cycle costings, are not included in most of the reviewed LCAs of MSW, they remain relatively common. A financial life cycle costing is present in eight of the reviewed LCAs, while an economic valuation of the environmental impacts is observed in five. Paper 1 also addresses decisions pertaining to the impact assessment stage of life cycle assessment. More than one quarter of the reviewed LCAs supply single score weighted results for the overall environmental performance of MSW management scenarios, although weighting is discouraged in the ISO guidelines (ISO 2006). The most common impact categories for which data are displayed include acidification potential (19/20), global warming potential (19/20), eutrophication (14/20) and resource consumption (11/20). The relative popularities of each impact category can inadvertently skew the interpretation of the LCA results in favour of those scenarios which show lesser environmental effects under the more commonly used impact categories. The comparative analysis in Paper 1 also includes a statistical analysis of the net energy use (NEU), global warming potential (GWP) and acidification potential (AP) results of the reviewed LCAs. As measures of statistical dispersion, the interquartile ranges of the NEU, GWP and AP values are lowest for the landfilling (AP, NEU) and thermal treatment (GWP) scenarios. The results of these statistical comparisons indicate that thermal treatment scenarios tend to have a better environmental performance than landfilling, while the comparative results for mixed treatment scenarios are less obvious. Nevertheless, a comparison of the relative environmental performances of MSW 187

treatment scenario types within each study does not provide a clear confirmation or repudiation of the waste hierarchy. The findings from the first paper demonstrate the lack of methodological transparency within the reviewed LCAs, which makes the LCA results difficult to interpret, and hampers meaningful comparisons between the results. The paper also reveals the extent to which the ISO 14044 guidelines for LCA (ISO 2006) are not followed. The results of this comparative analysis perhaps point to the need for a review of the ISO guidelines and more stringent peer-review of LCA research submitted for publication.

6.1.2 The incorporation of waste prevention activities into life cycle assessments of municipal solid waste management systems: Methodological issues (Paper 2)

The second paper addresses the methodological issues associated with incorporating waste prevention activities into life cycle assessments of waste management systems. In this paper, an attributional method of evaluating LCAs of waste which incorporate WPAs is proposed. The proposed model, designated the Waste Management And Prevention (WasteMAP) LCA, is a hybrid of the traditional product and waste LCA. It allows for the comparison of functionally equivalent waste management scenarios in which different quantities of waste are collected and treated. This functional equivalence assumes that waste prevention takes place through dematerialization, so long as there is a product service associated with the generated waste. Waste prevention activities through dematerialization are analogous to waste management techniques such as landfilling in that they do not affect the functional output (product services) of MSW-generating product systems (e.g., packaging for wines and spirits). The WasteMAP LCA model can be used to identify, in the results, the impacts and avoided impacts attributed to waste prevention, recycling, biological and thermal treatments, as well as landfilling, within a particular MSW management system. In the case of waste prevention through dematerialization, it is necessary to introduce an additional type of functional unit. Primary and secondary functional units are necessary 188

to ensure that all of the residential waste management scenarios that are compared with one another manage the same amount of waste and produce identical reference flows of functionally equivalent product services. The WasteMAP LCA model is suitable for LCAs of waste management systems that take into account multiple types of WPAs. The mathematical equations in Paper 2 describe how the results of an LCA using the WasteMAP LCA model would differ from the results obtained from a traditional LCA of MSW. Since the “cradle” of the traditional LCA of MSW is the moment of waste generation, this type of LCA lacks the upstream component necessary for the depiction of the WPAs taken into account with WasteMAP.

6.1.3 Life cycle assessments of wine and spirit packaging at the product and the municipal scale: A Toronto, Canada case study (Paper 3)

The third paper applies LCA to wine and spirit packaging systems both at a product scale (i.e., a one litre package for wines; a 750 ml package for spirits) and at a municipal scale (i.e., packaging for wines and spirits supplied to residents of the City of Toronto in 2008). At the scale of the individual package, the net environmental impacts do not take into account the effects of product consumption levels and the cumulative impacts from use of multiple types of products with the same functional output. The municipal scale does take these impacts into account, although the LCA results at this scale cannot be used to derive the environmental performance of individual packaging systems. The municipal scale LCA is a necessary intermediate step in order to depict WPAs within WasteMAP LCA system boundaries. The results of the individual package LCAs of Paper 3 reveal the superior environmental performances of the refillable glass container and of the aseptic carton in comparison to their packaging alternatives. While endpoint impact reductions from conventional single use glass containers can reach 87%, such reduction levels are not plausible when the system boundary is scaled up to incorporate all wine and spirit consumption at a municipal scale, for a variety of reasons. Examples of the causes of impact reduction limitations include the unsuitability of aseptic cartons (ACs) for the packaging of spirits and sparkling wines, the inability to use ACs and PET containers for 189

the aging of wine, and the infeasibility of returning refillable bottles to far away markets for reuse. Limitations such as these place a ceiling on the net environmental benefits that can be obtained through packaging substitutions. In contrast to the product scale LCAs, the municipal scale LCAs account for the actual levels of consumption of different packages supplying wines and spirits, along with a plausible alternative scenario substituting lighter weight and refillable containers that would provide a functionally equivalent packaging service. At this scale, the alternative packaging scenario produces endpoint level impact reductions of up to 42%, relative to the impacts from the 2008 reference.

6.1.4 Waste prevention and life cycle assessment of residential waste management in Toronto, Canada (Paper 4)

Paper 4 is a case study of the WasteMAP LCA model introduced in Paper 2. This model is applied to the residential waste management system of the City of Toronto, Canada, with scenarios incorporating six types of waste prevention activities. The LCA results are based on the evaluation and comparison of the net environmental impacts associated with various residential waste management scenarios. This LCA requires the collection and analysis of considerable amounts of primary data, as well as the use of LCA software (SimaPro 7.2), emissions databases and impact assessment methods. The residential waste management scenarios differ in the amount and type of waste prevention activity undertaken within the waste management system. Nevertheless, each scenario is considered functionally equivalent because the primary and secondary functional units remain consistent. The results of this LCA reveal the potential of various WPAs in reducing the net environmental impacts associated with the City of Toronto’s residential waste management system. While the amount of waste prevention undertaken under the waste prevention scenario is 3.5% of the quantity (by mass) of waste generated under the reference scenario, the changes in the net environmental impacts from the reference tend to be much larger. Three different LCIA methods (ReCiPe, IMPACT 2002+ and TRACI 2) are applied in order to examine the variation and consistency of the LCA results. 190

Although at times there are considerable differences in the results generated using the three methods, explanations for these differences are readily apparent from study of the modelling decisions and designs specific to each LCIA method. The results from Paper 4 also highlight the following: (1) the significance of the decision to account for recycled content when modelling waste prevention; (2) the related decision of whether or not to assume that prevented recycled material inputs would be reallocated to displace the virgin content of the products that continue to be produced; and (3) the additional benefits that can be accrued by targeting WPAs to products with low recycled content due to the significance of recycled material feedstock losses in diminishing the gains from waste prevention. The transportation component of the life cycle, sometimes dismissed in published LCAs of MSW (e.g., Chaya and Gheewala 2007; Mendes et al. 2004), can comprise a sizable proportion of the life cycle impacts. The waste collection and transportation components of the life cycle of the City of Toronto’s waste management system are considerable, as are the benefits that ensue from using a landfill (i.e., the Green Lane landfill) closer to the location of waste generation.

6.2 Policy implications and future research trajectories

The models and results from the four papers have implications for different audiences, such as LCA scientists/developers and practitioners, as well as policy-makers. For LCA scientists, the findings of the comparative analysis from Paper 1 suggest a widespread lack of methodological transparency in published LCAs from 2002 to 2008. These results indicate a need for more stringent LCA peer-review and the publication of detailed explanations as appendicies to published papers. The WasteMAP conceptual model supplies the LCA scientist and practitioner with a means of facilitating comparisons of multiple types of waste management options which include waste prevention and treatment activities. WasteMAP could be used to evaluate the hierarchy of waste management, identify those potential impacts attributable to waste prevention activities, and, in an LCA of waste, consider waste prevention as a waste management technique that is functionally equivalent to a form of waste treatment. 191

Policy-makers, including waste managers, responsible for minimizing the environmental impacts of waste management systems, should be especially interested in the findings of Paper 4. The findings could serve to reinforce the original justification for policies promoting various forms of waste prevention (e.g., European Commission 2005), as well as to provide encouragement for those considering the implementation of waste prevention policies. Perhaps of equal importance is the familiarity of policy-makers, as well as LCA practitioners, with an LCA tool which can be used to evaluate the benefits of including waste prevention activities in MSW management systems. For example, the system boundaries and functional units of the WasteMAP LCA model can be adopted by municipalities that are considering a specific array of WPAs for implementation in order to reduce their waste treatment costs and environmental impacts. The results of Paper 3 should also be of interest to policy-makers whose primary areas of focus are packaging and waste management. Although there is substantial uncertainty associated with the absolute endpoint level impacts of the various packaging systems evaluated (which is true for any LCA), one can have additional confidence in the estimates of packaging system impacts relative to one another. The potential endpoint level environmental impacts of refillable containers and aseptic cartons were found to be lower than those of other types of containers. The results also reveal the prominence of the transportation component of the life cycle, and perhaps surprisingly, that of secondary packaging for the RFG container life cycle. From a packaging policy perspective, the introduction of container reuse for domestic wines and spirits in Toronto, Canada could significantly reduce potential environmental impacts.

6.2.1 Transparency and consistency of published LCAs of MSW

Paper 1 reveals the extent to which many published LCAs of waste fail to disclose their methodological choices and assumptions. Very few of the reviewed LCAs of MSW identify or provide access to this critical information. This lack of transparency makes it difficult for waste managers, policy-makers and others to interpret and compare LCA results. 192

Due to the substantial amount of information required when defining the scope of an LCA, its data inputs and its assumptions, strict word limits in journal articles might have played a role in curtailing the inclusion of significant information. However, with the more common availability of online supplemental documentation, those who carry out LCAs increasingly possess opportunities to supply the information necessary in order to meet the ISO guidelines for LCA. These opportunities should be seized. The lack of consistency among the reviewed LCAs of MSW is clearly an issue, although not necessarily problematic. LCAs have different objectives and their system boundaries, assumptions, and data inputs must be tailored to meet these objectives. Yet, in the comparative analysis of Paper 1, there are examples of published LCAs of MSW where questionable and unusual choices, including certain system boundary curtailments (e.g., the omission of waste transportation from the life cycle), are poorly justified or left unjustified. A convergence in the adoption of particular assumptions that would be more representative of MSW management systems would facilitate the comparison of the results. In light of the findings of Paper 1, the methodological transparency necessary for the independent verification of results is addressed through the inclusion of detailed appendices to Papers 3 and 4, which would facilitate this verification process.

6.2.2 Comparison of environmental impacts of residential waste management systems that incorporate waste prevention activities

LCA results using WasteMAP can be used to identify net impacts and trade-offs (e.g., waste prevention versus recycling) within life cycles of residential waste management. The endpoint level impact results from Paper 4 are consistent with the waste hierarchy assumption that waste prevention activities display a better environmental performance relative to other waste management options. However, there remain numerous examples of waste prevention activities yet to be studied. Municipal solid waste managers can focus on the reduction of per capita waste generation by addressing decisions made in advance of the ‘end of life’ stage of waste management. For those municipalities, such as Toronto, Canada, which have initiated 193 policies to promote waste prevention, the WasteMAP model of undertaking an LCA of MSW could be used to provide insight into the net environmental burdens of their MSW management systems. However, many WPAs, such as the substitution of lightweight and refillable wine/spirit packaging systems, cannot be implemented by MSW managers alone. Effective policies and initiatives can nevertheless be introduced to encourage municipal residents and commercial enterprises to reduce the quantities of waste that they generate. Taking account of the lively debate within policy and academic circles on the merits of the cap and trade scheme for mitigating greenhouse gas emissions (e.g., Hanemann 2010), it may behoove policy-makers to consider a program that would implement an annual limit to the quantity and/or toxicity of MSW collected for treatment in a municipality or region. In these circumstances, the WasteMAP LCA could be a useful tool for evaluating the net impacts of such a measure. Since mass and toxicity are not adequate proxy indicators of the environmental impacts of a waste management system, it would be advisable to use LCA to evaluate the environmental performance of MSW management under the cap vs. the “business as usual” case. Such an LCA would need to address waste prevention (an aggregation of several waste prevention initiatives) as functionally equivalent to the various forms of waste treatment. For example, the cap could allow or encourage those institutions/producers/consumers capable of undertaking waste prevention activities to submit their proposals for inclusion in a waste prevention registry of initiatives for implementation. An initiative which could qualify under such a scheme is the Liquor Control Board of Ontario’s specific goal to reduce wine and spirits packaging waste generation by 10 000 tonnes annually by encouraging producers to reduce bottle weights and use alternatives to glass (LCBO 2006).

6.2.3 Effects of waste prevention activities on the management of residual waste in the MSW management system

The changing composition of waste has had significant effects on recycling systems. Those waste materials which negatively affect the sorting and processing of recyclables have been deemed “problematic materials” by waste managers, an example of 194

which is polylactide (PLA) biodegradable plastics (Solid Waste Management & Recycling 2008). Elevated contamination rates at sorting facilities and recycling processors have increased the costs of recycling, and decreased the quality of the recycled material (Lantz 2008). In light of such contamination and processing efficiency problems, a subsequent research application of the WasteMAP LCA could be the evaluation of the effects of the targeted waste prevention of problematic materials on the environmental performance of MSW treatment systems. This form of LCA would require the use of Equation 6 of WasteMAP, which was defined in Paper 2. The results from such an LCA could exemplify how changing the composition of the collected waste, via a WPA, has significant impacts on processes that would have been considered outside of the conventional product system boundaries associated with the waste materials added and removed. This type of evaluation could be used to aid in the development of product and packaging policies, specifically those targeting the generation of various problematic materials.

6.3 Concluding remarks

The production of LCAs of waste management has been guided and facilitated by clarifications in LCA methodology as well as by the introduction of detailed LCA process and emission databases. In spite of these clarifications, many published LCAs of MSW have not supplied input data in detail sufficient for independent verification of results. Results have also tended to omit the effects of activities undertaken to reduce per capita waste generation. The waste hierarchy has heretofore lacked adequate scientific backing via life cycle assessment, although, through its simplicity, it acts as a valuable rule of thumb for identifying the relative environmental performance of waste management options. Publications in the LCA scientific literature, such as those by Ekvall et al. (2007) and Gheewala (2009), have identified a need for LCAs of waste which incorporate waste prevention. The WasteMAP LCA is an attempt to address both of these issues, and to facilitate the design and application of future LCAs of waste management. 195

The LCAs in Papers 3 and 4 evaluate waste prevention at the product level and at the level of the MSW management system; and supply, in the text and appendices, the input data and assumptions required for independent result verification. The former paper addresses the net environmental impacts of several types of wine and spirit packaging systems, each packaging type responsible for different quantities of waste per unit volume packaged. In contrast, Paper 4 addresses the MSW management system of the City of Toronto and applies the WasteMAP model to allow for the comparison, on a functionally equivalent basis, of the LCA results of a reference scenario, based on 2008 data, with a scenario incorporating six types of waste prevention activities. Unlike Paper 3, this paper includes an evaluation of LCA results using three life cycle impact assessment methods (ReCiPe, IMPACT 2002+ and TRACI 2) in order to address the variation in the results. How can one evaluate the environmental burdens of a waste management system that incorporates waste prevention activities? By accepting an expansive depiction of MSW management, waste prevention can be considered a management activity that is functionally equivalent to waste treatments such as landfilling, incineration and recycling.

6.4 References

Ekvall T, Assefa G, Bjorklund A, Eriksson O, Finnveden G. 2007. What life-cycle assessment does and does not do in assessments of waste management. Waste Manage 27(8):989-996

European Commission. 2005. Impact assessment on the thematic strategy on the prevention and recycling of waste. COM 666 final.

Gheewala S. 2009. LCA of waste management systems—research opportunities. Int J LCA 14:589-590

Hanemann, M. 2010. Cap-and-trade: a sufficient or necessary condition for emission reduction? Oxford Review of Economic Policy 26 (2):225-252

ISO. 2006. Environmental Management – Life Cycle Assessment. Principles and framework. ISO 14040.

Lantz D. 2008. Mixed Results. Resour Recycl 11-15

196

Liquor Control Board of Ontario (LCBO). 2006. LCBO Annual Report 2005-2006. Toronto, Ontario: LCBO Corporate Communications. 46 pp.

Mendes MR, Aramaki T, Hanaki K. 2004. Comparison of the environmental impact of incineration and landfilling in Sao Paulo City as determined by LCA. Resour Conserv Recycl 41:47–63.

Solid Waste Management & Recycling (2008) The Dirty Dozen: Twelve materials that create problems for recycling plants. Solid Waste Manage & Recycl 13(6):51-52

197

Appendix 1

Glossary

198

Avoided Burden Allocation The “avoided burden” approach subtracts environmental emissions of the product systems targeted for prevention from the emissions of the treatment system for the MSW. The primary methodological complication of this approach is evident in cases in which the product system targeted for prevention has significant co-products, possibly resulting in the introduction of an asymmetry. For waste prevention through dematerialization, the avoided burdens (upstream and downstream) of the product systems targeted for prevention must be taken into account.

Attributional LCA An attributional LCA is used to describe a system and its environmental exchanges (Rebitzer et al. 2004) within a pre-set system boundary. It applies average emissions factors (e.g., the mean carbon dioxide emissions from an electricity grid) whereas the consequential approach applies marginal data (e.g., the emissions from the marginal source of electricity).

Consequential LCA A consequential LCA is “a model of causal relationships originating at the decision at hand” (Ekvall and Weidema 2004), addressing the economy-wide effects of a change in the functional outputs and inputs on material and energy flows to and from the environment (Curran et al. 2005). Unlike the attributional LCA, this method addresses the marginal effect of a change. The functional unit of a consequential LCA of waste prevention would be the amount of waste prevention one intends to undertake.

Co-product A co-product is defined as “any two or more products coming from the same unit process or product system” (ISO 2006).

Cut-Off Allocation In LCAs of MSW, it is common practice to apply the “cut-off” allocation, which excludes from the system boundary all products and emissions from waste management processes, such as heat, electricity, recycled material and compost.

Dematerialization The “process of fulfilling society’s functions with a decreasing use of material resources over time” (Van Der Voet et al. 2004).

Downstream Functional Unit The downstream functional unit of the WasteMAP case study is “mass of the City of Toronto’s residential waste collected and treated per year."

Functional Equivalence A qualitative description of the extent to which product services, accounted for through WasteMAP’s secondary functional unit, have similar properties. In addition, the upstream and downstream primary functional units are considered equivalent in that they quantify the management of waste through either treatment or prevention and reuse. 199

Functional Unit A reference unit to which the input and output data are normalized (ISO 2006). Examples include: “tonnes of residential waste generated by the residents of a municipality in a particular year” or “one litre of packaged wine consumed”

Hierarchy of Waste Management The hierarchy of waste management broadly depicts the options to manage waste, by order of preference (Price and Joseph 2000).

Life Cycle Assessment (LCA) A method to compile and evaluate the inputs, outputs and potential environmental impacts of a product and/or waste management system during its life cycle (ISO 2006).

Life Cycle Assessment of Waste A traditional LCA of MSW can be used to evaluate the life cycle impacts of a waste treatment system in which a particular waste stream is eliminated or reduced in size due to a waste prevention activity (WPA). It does not account for the net upstream impacts from implementing the WPA, nor the possible substitute product system(s) necessary to maintain an equivalent level of product services to the population.

Life Cycle Impact Assessment (LCIA) A tool for the evaluation of a product/waste management system using indicators for various impact categories that take into account life cycle inventory results (ISO 2000).

Primary Functional Unit The amount (mass or volume) of material addressed by the MSW management system on an annual basis.

Product Life Cycle Assessment The traditional product LCA is attributional, including the environmental exchanges of those processes within the defined product system. It includes the waste management stage only for the product studied and does not account for the potential effects of the removal or reduction in size of a waste stream on the treatment of the waste remaining in the MSW management system.

Product System A “collection of materially and energetically connected unit processes which performs one or more defined functions.” (ISO 1997)

Reference Flow A reference flow is a quantified amount of the product, including its parts, required so that a specific product system can deliver the performance depicted by the functional unit. The reference flow for composite products such as aseptic cartons is typically 200

identical to its list of composite materials, multiplied by a factor to scale it to the functional unit (Weidema et al. 2004).

Residential Waste Sold waste generated by households.

Secondary Functional Unit (SFU) An SFU is a functional unit that depicts the function supplied by a product service. It is used to describe the substitution ratio between the WPA and the replaced product service. It differs from the upstream functional unit in that it depicts the amount of a product service provided instead of the mass of waste left out of the waste stream.

System boundary Defined as the “interface between a product system and the environment or other product systems” (ISO 1997), or more specifically as a “set of criteria specifying which unit processes are a part of the product system” (ISO 2006).

Unit process The “smallest element considered in the life cycle inventory analysis for which input and output data are quantified” (ISO 2006).

Upstream Functional Unit The upstream functional unit of WasteMAP is "tonnes of city's products and materials left out of the waste stream through waste prevention and product reuse activities per year."

Waste Management The control of waste-related activities, including waste prevention and product reuse, with the aims of protecting the environment and human health, as well as resource conservation (adaptation of definition from Pongrácz and Pohjola (2004)).

Waste Management and Prevention (WasteMAP) Life Cycle Assessment The WasteMAP LCA is a hybrid of traditional product and waste LCAs which incorporates waste prevention activities (WPAs) into LCAs of waste management systems. This conceptual model allows the LCA practitioner to compare various waste management scenarios which allocate different quantities of waste for treatment, without violating the ISO 14044 (2006) international standard which states that LCA scenarios subject to different functional units should not be compared. The WasteMAP LCA treats waste prevention (including product reuse) activities as functionally equivalent to landfilling, recycling, thermal and biological treatments as means of managing waste. This functional equivalence necessitates that waste prevention takes place through dematerialization because the alternative waste management techniques (e.g., landfilling) do not affect the composition and magnitude of product services supplied to the population by waste-generating product systems (e.g., packaging for wine). WasteMAP is an attributional LCA that addresses waste prevention as an intrinsic part of a waste management system.

201

Waste Prevention Activity (WPA) A management process that is undertaken to reduce the amount of waste generated or collected.

Waste Treatment The landfilling, recycling, biological or thermal treatment of solid waste.

Waste The output from a product system that is for disposal.

References

Curran M, Mann M, Norris G. 2005. The international workshop on electricity data for life cycle inventories. J Clean Prod 13:853-862.

Ekvall T, Weidema B. 2004. System Boundaries and Input Data in Consequential Life Cycle Inventory Analysis. Int J LCA 9(3):161-171.

ISO (International Organization of Standardization). 2006. Environmental management – life cycle assessment – requirements and guidelines. ISO 14044.

ISO. 1997. Environmental Management – Life Cycle Assessment. Principles and framework. ISO 14040.

Van Der Voet E, Van Oers L, Nikolic I. 2004. Dematerialization: Not just a matter of weight. J Ind Ecol 8(4):121-137.

Weidema B, Wenzel H, Petersen C, Hansen K. 2004. The Product, Functional Unit and Reference Flows in LCA. Danish Environmental Protection Agency, Environmental News No. 70, 2004.

202

Appendix 2

Paper 1 – Supplemental Tables

203

Table A2.1 Primary functional units and system boundaries of reviewed LCAs of MSW

Source Explicit definition of Inclusion within system boundaries primary functional MSW transport Life cycle emissions unit of capital / infrastructure Arena et al. 2003 Yes Yes NM/U Aye and Widjaya 2006 Yes Yes No Beigl and Salhofer 2004 No Yes NM/U Buttol et al. 2007 No Yes Yes Chaya and Gheewala 2007 Yes No No Consonni et al. 2005 [a, b] Yes Yes Yes Di Maria and Fantozzi 2004 No Yes NM/U Emery et al. 2007 No Yes NM/U Eriksson et al. 2005 Yes Yes No Finnveden et al. 2005 / Moberg et al. 2005 Yes Yes No Hong et al. 2006 Yes No NM/U Kirkeby et al. 2006 No Yes NM/U Mendes et al. 2004 Yes No Yes Morris 2005 No Yes NM/U Özeler et al. 2006 Yes Yes NM/U Reich 2005 No Yes NM/U Rodriguez-Iglesias et al. 2003 Yes Yes NM/U Shmelev and Powell 2006 No Yes No Solano et al. 2002 [a, b] No Yes No Tan and Khoo 2006 Yes Yes NM/U Acronym NM/U–not mentioned / unclear

Table A2.2 Impact categories addressed in life cycle impact assessments and weighted valuations of impacts

Source Weighted Impact categories valuation s of AP GWP Eu RC HT PO Et SO H Other impacts impacts t x M Arena et al. No X X X Landfill volume, water 2003 pollution Aye and No X X X X Widjaya 2006 Beigl and No X X X Salhofer 2004 Buttol et al. No X X X X X X Volatile organic carbon 2007 Chaya and No X X X X X X X Gheewala 2007 Consonni et al. No X X X X Landfill volume 2005 [a, b] Di Maria and No X X X X X X X Fantozzi 2004 Emery et al. No X X X X X 2007 Eriksson et al. No X X X X X Volatile organic carbon, 2005 NOx Finnveden et al. Yes X X X X X X X X Non-treated waste, 2005 / Moberg NOx, SOx, NH3 et al. 2005 Hong et al. Yes X X X 2006 Kirkeby et al. No X X X X X X X X 2006 204

Mendes et al. No X X X 2004 Morris 2005 Yes X X X X X X Özeler et al. No X X X X X 2006 Reich 2005 Yes X X X X X NOx Rodriguez- Yes X X X X X X Carcinogens Iglesias et al. 2003 Shmelev and No Index of environmental Powell 2006 damage Tan and Khoo Yes X X X X 2006 TOTAL OUT 19 19 14 11 8 8 6 5 4 OF 19 Acronyms AP-Acidification potential; GWP-Global warming potential; Eut-Eutrophication; RC-Resource consumption; HT- Human toxicity; PO-Photochemical oxidants or ozone; Etx-Ecotoxicity; SO-Stratospheric ozone; and HM-Heavy metals.

Table A2.3 Economic impacts, sensitivity analysis, and electricity networks addressed

Source Financial costs Economic Sensitivity Location of Average or valuation of analysis electricity grid marginal source of environmental electricity impacts Arena et al. 2003 No No Not mentioned Italy (same) Average explicitly Aye and Widjaya Yes Yes (only for Yes Indonesia (same) Average 2006 -including capital GWP) Beigl and Salhofer Yes No Yes Different country Not mentioned 2004 -excluding capital Buttol et al. 2007 No No No Italy (same) Average Chaya and No No Not mentioned Thailand (same) Average Gheewala 2007 explicitly Consonni et al. Yes No No Italy (same) Marginal 2005 [a, b] -including capital Di Maria and No No No Not mentioned Not mentioned Fantozzi 2004 Emery et al. 2007 Yes No Not mentioned United Kingdom Not mentioned -excluding capital explicitly (same) Eriksson et al. 2005 Yes Yes Not mentioned Sweden (same) Marginal -excluding capital explicitly Finnveden et al. No Yes Not mentioned EU Marginal 2005 / explicitly Moberg et al. 2005 Hong et al. 2006 No No No Not mentioned Not mentioned Kirkeby et al. 2006 No No No Denmark (same) Marginal (coal) Mendes et al. 2004 No No Yes Brazil (same) Average Morris 2005 No Yes (certain No Not mentioned Marginal (coal) impacts) Özeler et al. 2006 No No No Not mentioned Not mentioned Reich 2005 Yes Yes No Not mentioned Marginal (coal) -including capital Rodriguez-Iglesias No No Not mentioned Not mentioned Not mentioned et al. 2003 explicitly Shmelev and Yes No Yes Not mentioned Not mentioned Powell 2006 -including capital Solano et al. 2002 Yes No No Not mentioned Unclear (claimed [a, b] -excluding capital both avg. and marg.) Tan and Khoo 2006 No No No Singapore (same) Average

205

Appendix 3

Paper 3

206

A3.1 Wine and sprit container, closure and capsule data, assumptions and calculations

A3.1.1 Containers

A3.1.1.1 Description of 2008 LCBO wine and spirit volume sales datasets

The 2008 wine and spirits volume sales datasets include all of the wine and spirit sales classified by container type and volume, type of closure, and product brand. The information used from this dataset is as follows: (1) the percentage of 2008 wine/spirit volume sales in glass bottles, PET bottles, aseptic cartons and bag-in-box containers; (2) the percentage of total wine/spirit volume sales by container size; and (3) the percentage of wine and spirit containers using various types of closures (corks, synthetic corks, metal and plastic screw caps, pull tabs, pressure tabs, pour spouts and other types of closures). Brand information was only used in the calculation of the estimated mass of wine and spirit containers, without closure or label, by the size and type of container (Table A3.1). Data in Table 3.1 are reported in aggregate form rather than by brand due to confidentiality.

A3.1.1.2 Procedure used to exclude wine and spirit coolers from the LCBO datasets

As discussed in Paper 3, the alternative wine and spirit packaging options do not pertain to ciders or wine and spirit coolers. Therefore, it was necessary to exclude the effects of these types of beverages from the LCA data inputs. After in depth review of the spirit sales volume dataset from the LCBO, it was determined that sales of spirit coolers could be eliminated from the spirits dataset by omitting sales of 30 ml, 120 ml, 200 ml (6 brands), 250 ml, 260 ml, 270 ml, 275 ml, 300 ml (one example), 330ml, 333 ml, 341 ml, 355 ml, 400 ml (5 brands), 473 ml (2 exceptions), 475 ml, 500 ml (4 brands), 650 ml, 700 ml (one brand), 710 ml, 900 ml (one brand), 1000 ml, 1065 ml, 1080 ml, 1100 ml, 1200 ml, 1320 ml, 1332 ml, 1360 ml, 1364 ml, 1420 ml, 1600 ml, 1620 ml, 2000 ml, 2046 ml, 2130 ml, 2728 ml, 3960 ml, 4092 ml, 4260 ml, 5676 ml, 7920 ml, and 207

8520 ml containers (or sets of containers). Similarly, wine coolers and ciders were removed from the wine volume sales dataset by omitting sales of 1364 ml and 2046 ml sets of containers, as well as by individually identifying and removing those remaining cooler and cider beverage sales. These procedures result in the exclusion of approximately 35.5% of spirit sales, and 2.3% of wine sales (by volume).

A3.1.1.3 Mass of wine and spirit containers, without closure or label, by size and type of container

Estimates of the average mass of each type of empty wine/spirit container supplied to the residents of Toronto were generated via the following procedure: (1) a sample of 315 bestselling wine/spirit products (representing 33.1% and 55.1% of 2008 wine and spirit sales in Ontario by volume, respectively), including the container, its contents, the closure, capsule and label, were weighed at an LCBO outlet using a digital kitchen scale; (2) the LCBO 2008 volume sales dataset was used to determine the degree of representation of the measured sample, comprising 165 wines and 150 spirits among the top 500 most popular products, for each type of container sold, by volume of the container; and (3) the estimated masses of the liquid contents, the closures, capsules and labels were subtracted from the measured mass values of the products (see Table A3.1). For the 750 ml PET wine bottle and 1 litre aseptic carton, the estimates were based on data from published LCAs. For the 1 litre PET wine bottle, it was based on the mass measurement of an empty bottle used to package 90.2% of the volume sales in that type of bottle.

Table A3.1 Estimated mass of wine and spirit containers, without closure or label, by size and type of container

Volume of Glass bottle PET bottle Aseptic carton container (ml) Mass (g), Sample % Mass (g), Sample % Mass (g) Sample % weighted Representation stand. dev. Representation and Representation stand. dev.1 by Vol. Sales and by Vol. Sales container by Vol. Sales and container mass/litre container mass/litre mass/litre Wine 750 502±62 38.8 54 N/A (WRAP N/A N/A (669 g/L) (n=93/3244 (72 g/L) 2008, Franklin brands) Associates 2006) 208

1000 543±65 30.4 58 90.2 (based on 34.5 (tetra N/A (Franklin (543 g/L) (n=17/54 brands) (58 g/L) mass prisma) Associates 2006) measurement of (34.5 g/L) an empty bottle) 1500 740±49 69.7 N/A N/A 50 (tetra 100% (493 g/L) (n=48/298 brik) (1/1 brand) brands) (33 g/L) 2000 692±33 68.3 N/A N/A N/A N/A (346 g/L) (n=4/8 brands) 3000 1156±1 99.9 N/A N/A N/A N/A (385 g/L) (n=2/27 brands) Spirits 200 213±23 37.1 29 99.0 N/A N/A (1065 g/L) (n=12/144 (145 g/L) (1/2 brands) brands) 375 335±34 21.8 41±3 54.8 N/A N/A (893 g/L) (n=16/259 (109 g/L) (8/32 brands) brands) 750 515±85 47.0 63±13 90.9 N/A N/A (687 g/L) (n=49/1976 (84 g/L) (4/9 brands) brands) 1140 665±86 69.2 71±5 78.7 N/A N/A (583 g/L) (n=38/265 (62 g/L) (2/6 brands) brands) 1750 1,211±147 36.1 87±12 67.6 N/A N/A (692 g/L) (n=9/118 brands) (50 g/L) (11/28 brands) 1 The weighted standard deviation for each type of container was calculated as follows, assuming that Ax:Ay is a list of the percent representation of each container mass sample of wines or spirits of the particular container size and type, by volume of sales in 2008 (e.g., Ax would equal 5% if it represented one brand of spirit comprising 5% of the sample of 750 ml spirit containers, by volume of sales in 2008); and assuming that Bx:By are the respective masses of the sampled containers (excluding the liquid contents and secondary packaging).

The average mass (AvgM) of each container is calculated as follows: AvgM = ( (Ax:Ay*Bx:By)) / ((Ax:Ay))

The variance (VarM) of the masses is calculated as follows: VarM = ( (Ax:Ay*(Bx:By-AvgM)²)) / ((Ax:Ay)-1)

The equation for calculating the standard deviation is: = VarM

Mass measurements for 25 brands of spirits used as inputs for calculating the average mass of spirit containers are listed in Table A3.2.

Table A3.2 Measured densities of popular spirits purchased in Ontario, used as inputs for calculating the average mass of spirit containers

Brand Type of spirit Measured density (g/ml) Absolut Vodka 0.944 Alberta Premium Rye 0.952 Appleton Rum Special Rum 0.956 Bacardi Gold Rum 0.953 Bacardi Superior Rum 0.946 Bailey’s Irish Cream Coffee cream liqueur 1.070 Beefeater Gin 0.944 Bombay Saphire Gin 0.942 Canadian Club Premium Whiskey 0.956 Captain Morgan Dark Rum 0.939 Captain Morgan Gold Rum 0.956 209

Captain Morgan Spiced Rum 0.957 Crown Royal Whiskey 0.946 Disaronno Spirit 1.080 Fireball Cinnamon Whiskey 1.043 Grand Marnier Liqueur 1.028 J&B Whisky 0.948 Jack Daniel’s Whiskey 0.949 Jagermeister Liqueur 1.007 Kahlua Coffee Liqueur 1.136 Luxardo Sambuca 1.087 Smirnoff Vodka 0.950 Smirnoff Raspberry Vodka 0.968 St. Remi Brandy 0.953 Southern Comfort Spirit 0.994

For the remainder of the brands for which the densities were not measured, the following substitutions were employed (Table A3.3):

Table A3.3 Spirit density substitutions for brands with unmeasured densities

Brand Type of spirit Substitute brand for density Assumed density (g/ml) Absolut Mandrin Vodka Smirnoff Raspberry Twist Vodka 0.968 Alberta Pure Vodka Vodka Smirnoff 0.950 Appleton Estate V/X Rum Appleton Rum Special 0.956 Bacardi 1873 Rum Bacardi Gold Rum 0.953 Bacardi Black Rum Captain Morgan Dark 0.939 Baileys Caramel Irish Cream Cream Liqueur Bailey’s Irish Cream 1.070 Baileys Mint Chocolate Irish Cream Cream Liqueur Bailey’s Irish Cream 1.070 Ballantine’s Whisky J&B 0.948 Banff Ice Vodka Vodka Smirnoff 0.950 Black Velvet Whisky Canadian Club Premium 0.956 Canadian Club Classic Whisky Canadian Club Premium 0.956 Canadian Club Reserve Whisky Canadian Club Premium 0.956 Captain Morgan White Rum Rum Bacardi Superior 0.946 Finlandia Vodka Smirnoff 0.950 Gibson’s Finest Whisky Canadian Club Premium 0.956 Gibson's Finest Sterling Edition Whisky Canadian Club Premium 0.956 Golden Wedding Whisky Canadian Club Premium 0.956 Gordon’s Dry Gin Gin Beefeater 0.944 Grey Goose Vodka Vodka Smirnoff 0.950 Iceberg Vodka Vodka Smirnoff 0.950 Jim Beam White Label Bourbon Bourbon Jack Daniel’s 0.949 Johnnie Walker Red Label Whisky Canadian Club Premium 0.956 Lamb's White Rum Rum Bacardi Superior 0.946 Prince Igor Vodka Vodka Smirnoff 0.950 Russian Prince Vodka Smirnoff 0.950 Seagrams 83 Whisky Canadian Club Premium 0.956 Seagrams V.O. Whisky Canadian Club Premium 0.956 Skyy’s Vodka Vodka Smirnoff 0.950 Smirnoff Green Apple Twist Vodka Vodka Smirnoff Raspberry Twist Vodka 0.968 Smirnoff Lime Twist Vodka Vodka Smirnoff Raspberry Twist Vodka 0.968 Smirnoff Orange Twist Vodka Vodka Smirnoff Raspberry Twist Vodka 0.968 Smirnoff Raspberry Twist Vodka Vodka Smirnoff Raspberry Twist Vodka 0.968 Smirnoff Strawberry Twist Vodka Vodka Smirnoff Raspberry Twist Vodka 0.968 Stolichnaya Vodka Vodka Smirnoff 0.950 Tanqueray Dry Gin Gin Bombay Saphire 0.942 Wiser's Deluxe Whisky Canadian Club Premium 0.956 Wiser's Special Blend Whisky Canadian Club Premium 0.956 210

The density estimate of the container contents becomes more important as the ratio between the packaging mass and the mass of the container contents increases. Thus, the effect of wine and spirit density uncertainty on container mass is much larger for AC and PET containers than for glass. For AC and PET wine containers, these uncertainty issues for calculations using the 1g/ml wine density were avoided by using (1) AC and PET container masses cited from peer-reviewed sources; and (2) mass measurements of empty containers.

A3.1.1.4 Assumed mass of each type of lightweight single use glass container

As discussed in Paper 3, lightweight single use glass bottles were assumed to be 80% of the measured average mass of the conventional single use glass bottle – a conservative estimate relative to estimates in the literature (e.g., 365 g for a 750 ml wine bottle according to WRAP 2008). Table A3.4 lists the assumed mass of each lightweight wine and spirit glass container, by the size of the container.

Table A3.4 Assumed mass of lightweight single use glass bottles

Size of Container Assumed mass (grams)1 Wine 750 401 (CSU mass 501 g) 1000 434 (CSU mass 543 g) 1500 591 (CSU mass 739 g) 2000 554 (CSU mass 692 g) 3000 925 (CSU mass 1,156 g) Spirit 200 170 (CSU mass 213 g) 375 268 (CSU mass 335 g) 750 410 (CSU mass 515 g) 1140 532 (CSU mass 665 g) 1750 967 (CSU mass 1,211 g) 1 Lightweight single use glass bottles are assumed to be 80% of the mass of the average conventional single use glass bottle.

A3.1.1.5 Container mass equations

For the municipal scale LCA, the following equations are used to calculate the volume of each type of wine and spirit container, the number of containers sold in Toronto in 2008, and mass of each type of container:

211

Vts = Vs * %Conts

ConNts = Vts/Conv

ConMts = Nts*Conm

where Vts signifies the total volume (in litres) of wines/spirits sold in a particular type of container (e.g., one litre aseptic carton) in Toronto in 2008; Vs is the volume sold in

containers of a particular size (e.g., a one litre container); %Conts is the percentage of

containers of a particular size that are of a particular type (e.g., aseptic carton); ConNts is the number of containers of a particular size and type that were sold in Toronto in 2008;

Conv is the volume of the individual container; ConMts is the total mass of empty

containers of a particular size and type that were sold in Toronto in 2008, and Conm is the average mass of the individual container (from Tables A3.1 and A3.4)

A3.1.2 Closures

A3.1.2.1 Mass of closures, by type

Wine and spirit containers possess a wide variety of closures, the majority as metal screw caps, corks, and synthetic corks. The average mass of each type of closure was acquired from published LCAs by PricewaterhouseCoopers (PwC)/ECOBILAN (2008) and Franklin Associates (2006), closure mass measurements, and assumptions based on surrogates where data were not available (Table A3.5).

Table A3.5 Estimated mass of wine and spirit container closures

Type of closure Estimated mass Source (as identified in 2008 LCBO sales data) (grams) Cork 3.5 PwC/ ECOBILAN (2008) Cork composite 3.5 Assumed identical to cork Cork-t 3.5 Author’s estimate based on measurements (n=1) Metal lid 4.6 Assumed identical to metal screw cap Metal screw cap 4.6 PwC/ ECOBILAN (2008) Natural cork 3.5 PwC/ ECOBILAN (2008) None / N/A / Other-miscellaneous1 See note See note No twist crown steel 4.6 Assumed identical to metal screw cap Plastic lid 3 Assumed identical to plastic screw cap Plastic screw cap 3 Author’s estimate based on measurements Plastic screw cap for aseptic carton 2.2 Franklin Associates (2006) Pressure cap 9.5 Author’s estimate based on measurements (n=5) 212

Synthetic cork 6.2 PwC/ ECOBILAN (2008) Synthetic cork-t 6.2 Assumed identical to synthetic cork 1 It is assumed that those containers with closures identified in the LCBO dataset as “none” “N/A” and “other-miscellaneous” possess closures in identical proportions as the remaining containers. A random selection of these containers revealed that they do possess closures. The “none” designation is often associated with containers within gift boxes or other packaging.

A3.1.2.2 Material composition of closures, by type

Table A3.6 lists the estimated material composition of wine/spirit closures. There are two cases in which the closure material types were unknown and the author had to make an unsubstantiated assumption. In both of these instances, the assumptions would have a negligible effect on the results since they pertained to very uncommon closure types.

Table A3.6 Estimated material composition of closures

Type of closure % mass Assumed material type Source (as identified in 2008 LCBO sales data) composition1 Cork / Cork composite / Cork-t1 / Natural 100% Cork PwC/ ECOBILAN (2008), cork / Pressure cap except for Cork-t Cork-t 36% Cork Author’s estimate based on 64% HDPE measurements (n=1) Metal lid / No twist crown steel 100% Steel Author’s assumption in lieu of missing data Metal screw cap 89.9% Aluminium PwC/ ECOBILAN (2008) 7% Expanded PET 2% Tin 0.6% PVC 0.5% Kraft paper None / N/A N/A N/A (assumption of no N/A closure) Other-miscellaneous 50% Aluminum Author’s assumption in lieu of 25% PP missing data 12.5% HDPE 12.5% LDPE Plastic lid / Plastic screw cap / Plastic 50% HDPE Franklin Associates (2006) for screw cap for aseptic carton 50% PP plastic screw cap for aseptic carton Synthetic cork / Synthetic cork-t1 68% LDPE PwC/ ECOBILAN (2008) for 16% HDPE synthetic cork 16% PP 1 The top of the “cork-t” and “synthetic cork-t” can be composed of many materials, such as wood, metal, glass and plastic. Plastic tops appear to be the most common and are assumed to be representative of the category.

A3.1.2.3 Adoption levels of closures, by type

Table A3.7 lists the adoption levels of container closures types for wines and spirits sold by the LCBO in Ontario, by size and type of container (excluding bag-in- 213 box), in 2008. These figures are derived from the 2008 LCBO volume sales datasets. Using the data from Tables A3.5 to A3.7, it was possible to estimate the change in the mass and composition of the closures in the residential waste under each LCA scenario.

Table A3.7 Estimated adoption levels of container closures for wines and spirits sold by the LCBO in Ontario, by size and type of container, 2008

Volume of container (ml) % of containers with each closure type Glass container closures PET container closures Aseptic carton closures Wines 750 35.04% metal screw cap 98.99% metal screw cap N/A 34.62% cork 1.01% plastic screw cap 23.74% synthetic cork 4.56% pressure cap 0.98% cork-t 0.56% natural cork 0.41% plastic screw cap 0.05% no twist crown steel 0.04% synthetic cork-t 0.00% cork composite 0.00% other-miscellaneous 1000 80.77% metal screw cap 100.00% metal screw cap 100.00% plastic screw cap 5.43% cork 3.84% synthetic cork 3.44% plastic screw cap 3.02% lid metal 2.39% natural cork 1.11% cork-t 1500 46.14% metal screw cap N/A 100.00% plastic screw cap 22.37% cork 19.02% synthetic cork 6.41% natural cork 3.05% pressure cap 2.59% metal lid 0.18% cork-t 0.04% other-miscellaneous (99.8% total) 2000 68.71% metal screw cap N/A N/A 31.29% metal lid 3000 99.90% metal screw cap N/A N/A 0.09% cork 0.01% natural cork 0.01% pressure cap (100.01% total) Spirits 200 41.77% metal screw cap 98.96% plastic screw cap N/A 21.64% plastic screw cap 1.04% metal screw cap 18.47% metal lid 12.04% N/A 5.40% plastic lid 0.29% none 0.19% cork-t 0.18% other-miscellaneous 0.02% natural cork 0.01% synthetic cork 0.00% other - aluminum 0.00% synthetic cork-t (100.01% total) 375 46.03% metal screw cap 56.87% metal screw cap N/A 20.69% metal lid 41.11% plastic screw cap 214

16.59% plastic screw cap 2.02% N/A 7.56% plastic lid 4.56% N/A 4.56% cork-t 0.01% none 0.00% synthetic cork 750 36.15% metal screw cap 63.59% plastic screw cap N/A 25.66% plastic screw cap 35.36% metal screw cap 15.12% metal lid 1.05% N/A 9.49% plastic lid 8.90% cork-t 4.17% N/A 0.28% other-miscellaneous 0.17% synthetic cork-t 0.03% cork 0.00% other - aluminum 0.00% natural cork 0.00% turn up spout 0.00% pressure cap 0.00% none 0.00% synthetic cork 0.00% cork composite (99.97% total) 1140 38.07% metal screw cap 54.94% plastic screw cap N/A 24.01% plastic lid 45.06% N/A 17.34% plastic screw cap 16.48% metal lid 2.75% cork-t 1.35% N/A 0.00% other-miscellaneous 1750 31.14% metal screw cap 58.52% N/A N/A 28.71% N/A 36.35% plastic screw cap 18.83% plastic screw cap 4.03% metal screw cap 18.09% plastic lid 1.10% plastic lid 2.69% metal lid 0.35% cork-t 0.18% other – miscellaneous 0.00% cork (99.99% total)

A3.1.2.4 Closure mass equations

For the municipal scale LCA, the following equations are used to estimate the mass of each type of material in the container closures supplied in Toronto in 2008, for each type of container closure (CloMtsm):

CloNts = ConNts*%Clots

CloMtsm = CloNts*%Clom

where CloNts signifies the total number of closures of particular type (e.g., a cork) used on a particular type of container in Toronto in 2008; %Clots is the percentage of containers of a particular size and type that use a closure of a particular type; CloMtsm is 215

the mass of a particular closure material on containers of a particular size and type that

were sold in Toronto in 2008; and %Clom is the percentage of the mass of a particular type of closure that is composed of a particular material (from Tables A3.5, A3.6 and A3.7). For the individual packages, the equations are used to estimate the mass and material composition of the closures used for the 1 litre wine and 750 ml spirit container.

A3.1.3 Capsules

A3.1.3.1 Mass of capsules, by type

Wine and spirit containers using cork closures possess one of four types of capsules, including those composed of aluminum, polylaminate, polyvinylchloride (PVC) and tin. The average mass of each type of capsule was obtained from mass measurements undertaken by the author (Table A3.8).

Table A3.8 Estimated mass of wine and spirit container capsules

Type of capsule Mass (grams) # of measurements Aluminum 0.8 n=3 Polylaminate 1.3 n=5 Polyvinylchloride (PVC) 0.8 n=2 Tin 4.5 n=2

A3.1.3.2 Material composition of capsules, by type

The average material composition of PVC capsules was obtained from mass measurements undertaken by the author in which the aluminum in the top cap was removed and measured separately from the remainder of the PVC capsule (Table A3.9). Aluminum and tin capsules were assumed to be composed of the materials supplying their designations. The percent of the mass of polylaminate capsules composed of aluminum and LDPE was estimated using the following data: (1) the thickness of each material within the average wine capsule: 36.4% aluminum and 63.6% LDPE (Rivercap n.d.); (2) the mass measurement for polylaminate capsules (1.3 g); and (3) the densities of 216

aluminum (2.7 g/cm³ - Encyclopedia of Earth 2010) and LDPE (0.9175 g/cm³ - IDES n.d.).

Table A3.9 Estimated material composition of capsules

Type of capsule % mass Assumed material type composition Aluminum 100% Aluminium Polylaminate 62.7% Aluminium 37.3% Polyethylene Polyvinylchloride (PVC) 7% Aluminum (in top cap) 93% PVC Tin 100% Tin

A3.1.3.3 Adoption levels of capsules, by type

The estimated levels of use of each type of container capsule were based upon survey data from Wine Business Monthly’s 2010 Capsule Report (Fisher 2010) for the United States market (Table A3.10). Using the data from Tables’ A3.8 to A3.10, it was possible to estimate the change in the mass and composition of the capsules in the residential waste under each LCA scenario.

Table A3.10 Estimated adoption levels of wine container capsules for wine sold in Ontario, 2008

Type of capsule % of capsules1 Aluminum 5.6% Polylaminate 65.8% Polyvinylchloride (PVC) 20.4% Tin 8.2% 1 Percentages based upon the adoption levels of various types of wine capsules for wines costing between US$7-13.99 per bottle (Fisher 2010).

A3.1.3.4 Capsule mass equations

For the municipal scale LCA, the following equations are used to calculate the mass of each type of material in the container capsules supplied in Toronto in 2008, for

each type of container closure (CloMtsm):

CapNts = ConNts*%Capts

CapMtsm = CapNts*%Capm

217

where CapNts represents the total number of capsules of particular type (e.g., a cork) used on a particular type of container in Toronto in 2008; %Capts is the percentage of

containers of a particular size and type that use a capsule of a particular type; CapMtsm is the mass of a particular capsule material on containers of a particular size and type that were sold in Toronto in 2008; and %Capm is the percentage of the mass of a particular type of capsule that is composed of a particular material (from Tables A3.8, A3.9 and A3.10). For the individual packages, the equations are used to estimate the mass and material composition of the capsules used for the 1 litre wine and 750 ml spirit container.

A3.2 Wine/spirit packaging transportation data, assumptions and calculations

A3.2.1 Transportation of wine/spirit packages to filling facility

Weighted mean transport distances from the container manufacturer to the packager were based upon the responses to the questionnaires sent to wine and spirit companies (distance estimates: 611±1765 km by ship and 824±810 km by land). This transportation life cycle component takes into account the transportation of the container as well as the packaging materials that will become losses at the packaging facilities. Emissions from the transportation of closures, capsules and labels were considered negligible and thus excluded from the LCA scenarios. Due to confidentiality agreements with the questionnaire respondents, it was necessary to keep the locations of the container manufacturer and filling facilities secret. For those questionnaire responses identifying “GTA” (Greater Toronto Area) and “NE USA” (Northeastern United States of America) as the geographic origins of some of the containers, the estimated distances were based upon the assumption of the “GTA” as the City of Toronto, and “NE USA” as the City of Boston. It was assumed that 50% of the land-based transport occurred by rail, with the remainder by truck. The distances between international ports were obtained from the “World Ports Distances Calculator” available at http://www.distances.com/. Nautical miles were converted to km by multiplying by 1.852. For distances traveled using land-based transport, the first distance supplied through Google Maps was selected (Google 2010). 218

A3.2.2 Transportation of packaged wines/spirits from the filling facility to Toronto, Canada in 2008

Statistics Canada data on packaged wine and spirit imports (excluding coolers) to Ontario in 2008 were used to identify the countries of origin of these products. Wines and spirits not shipped in bulk were assumed to be packaged in their countries of origin. Bulk imported wines and spirits packaged in Ontario are considered "domestic" in the Statistics Canada, LCBO, Canadian Vintners Association (CVA) and the Association of Canadian Distillers (ACD) statistics. For example, wines "Cellared in Canada" are domestic. Only those transportation emissions attributed to the wine and spirit package, including the container, closure, capsule and label, are included within the system boundary. Those emissions resulting from the container contents are excluded. Statistics Canada data on the volume of wines and spirits imported to Ontario were used to identify the amount of total wine/spirit imports from each country of import (Table A3.11).

Table A3.11 Wine/spirit import markets and the percent of wine/spirits imports to Ontario, Canada from each in 2008

Bottled wine imports Bottled spirit imports Countries % of Total, by volume1 Countries % of Total, by volume1 Italy 24.70 United Kingdom 23.68 Australia 17.69 Russian Federation 12.60 France 15.77 France 12.04 US-California 11.93 Sweden 7.79 Chile 9.28 Italy 4.57 Argentina 4.05 US-Kentucky 4.35 South Africa 3.78 Ireland, Republic of (EIRE) 4.28 Spain 3.14 US-California 3.51 Portugal 2.86 US-Illinois 3.50 Germany 2.24 Mexico 3.36 New Zealand 1.72 US-Florida 2.72 Greece 0.50 US-Tennessee 2.49 Japan 0.43 Cuba 2.37 Hungary 0.28 Finland 2.24 US-New York 0.25 Poland 2.05 Israel 0.19 Jamaica 0.88 Montenegro 0.17 Germany 0.85 US-New Jersey 0.12 Barbados 0.70 US-Kentucky 0.12 US-Puerto Rico 0.69 Austria 0.10 US-Minnesota 0.54 Bulgaria 0.09 Greece 0.54 US-Washington, state 0.07 Switzerland 0.46 Mexico 0.07 South Africa 0.44 US-Oregon 0.04 Netherlands 0.35 Korea, South 0.04 US-Maryland 0.28 Romania 0.04 Guyana 0.27 219

Macedonia (FYROM) 0.03 Trinidad and Tobago 0.23 Jamaica 0.03 Portugal 0.20 United Kingdom 0.03 Croatia 0.19 Poland 0.03 Canada 0.19 Canada 0.02 Bermuda 0.19 Georgia 0.02 US-New Jersey 0.14 Ukraine 0.02 Denmark 0.14 Moldova, Republic of 0.01 Spain 0.14 US-Maryland 0.01 Montenegro 0.13 Slovenia 0.01 Ukraine 0.12 Lebanon 0.01 Hungary 0.09 US-Illinois 0.01 Dominican Republic 0.08 Brazil 0.01 Austria 0.08 Uruguay 0.01 Czech Republic 0.06 Switzerland 0.01 Moldova, Republic of 0.05 Denmark 0.01 Lebanon 0.04 Russian Federation 0.01 Iceland 0.04 US-Florida 0.01 US-Arkansas 0.04 China 0.00 US-U.S. Virgin Is. 0.04 Morocco 0.00 US-Ohio 0.03 Cyprus 0.00 US-Connecticut 0.03 US-Virginia 0.00 Armenia 0.02 Croatia 0.00 New Zealand 0.02 Thailand 0.00 Israel 0.02 United Arab Emirates 0.00 Columbia 0.02 Turkey 0.00 US-New York 0.02 US-North Carolina 0.00 United Arab Emirates 0.02 US-Other States 0.00 Australia 0.01 Luxembourg 0.00 Georgia 0.01 Netherlands 0.00 Cyprus 0.01 US-Minnesota 0.00 Peru 0.01 US-Texas 0.00 Albania 0.01 Belgium 0.00 Korea, South 0.01 US-Connecticut 0.00 Virgin Islands, British 0.01 Sweden 0.00 US-Idaho 0.00 India 0.00 US-Maine 0.00 US-Missouri 0.00 US-Texas 0.00 US-Idaho 0.00 Latvia 0.00 Antigua and Barbuda 0.00 Panama 0.00 US-Hawaii 0.00 Lithuania 0.00 Bolivia 0.00 Saint Kitts 0.00 US-Arkansas 0.00 US-Missouri 0.00 US-Michigan 0.00 Slovenia 0.00 Costa Rica 0.00 India 0.00 US-Pennsylvania 0.00 US-Other States 0.00 Albania 0.00 Belgium 0.00 Armenia 0.00 Saint Lucia 0.00 Bahamas 0.00 Turkey 0.00 Barbados 0.00 Brazil 0.00 Bermuda 0.00 US-Indiana 0.00 Cayman Islands 0.00 Guatemala 0.00 Columbia 0.00 Bahamas 0.00 Cuba 0.00 Estonia 0.00 Czech Republic 0.00 US-Georgia 0.00 Dominican Republic 0.00 Bulgaria 0.00 Estonia 0.00 Cayman Islands 0.00 Finland 0.00 Japan 0.00 Guatemala 0.00 Philippines 0.00 Guyana 0.00 US-Michigan 0.00 Vietnam 0.00 1 The volumes of wine/spirits imported from those import markets listed with a 0.00 percent market share are taken into account in the transport calculations. 220

The figures displayed in Table A3.11 were calculated from import volume data acquired using Statistics Canada (2008) Harmonized System (HS) codes for each type of wine and spirit import, excluding coolers, to Ontario, Canada in 2008. Tables A3.12 and A3.13 list the Harmonized System (HS) Codes selected to provide the quantity of packaged wines and spirits imported to Ontario, Canada in 2008.

Table A3.12 HS Codes selected to provide the quantity of packaged wine imported to Ontario, Canada in 2008

Harmonized System Type of packaged wine imported to Ontario, Canada in 2008 (HS) Code 2204101000 Sparkling wine, of an alc strt by vol <= 22.9% vol, incl champagne 2204109000 Other sparkling wine, of an alc strt > 22.9% vol, except champagne 2204211010 Icewine 2204211091 Grape wine, white, nes, alc strength by volume<=13.7% vol, in containers<=2 L 2204211092 Grape wine, red, alc strength by volume<=13.7% vol, in containers <=2 litres 2204211099 Grape wines, nes, alc strength by volume <=13.7% vol, in containers <=2 L 2204212100 Grape wines, nes,incl fort,alc strength by vol >13.7% vol<=14.9% vol, ctnr<=2 L 2204212200 Grape wines, nes, incl fort, alc strength by vol >14.9% vol<=15.9% vol, cntr<=2 L 2204212310 Sherry, alc strength by volume >15.9% vol<=16.9% vol, containers<=2 litres 2204212390 Grape wines, nes, incl fort, alc strength by vol >15.9% vol<=16.9% vol,ctnr<=2 L 2204212410 Sherry, alc strength by volume >16.9% vol<=17.9% vol, cntr<=2 litres 2204212420 Madeira, alc strength by volume >16.9% vol<=17.9% vol, ctnr<=2 litres 2204212490 Grape wines,nes,incl fort, alc strength by vol >16.9% vol<=17.9% vol, ctnr<=2 L 2204212510 Sherry, alc strength by volume >17.9% vol<=18.9% vol, ctnr<=2 litres 2204212520 Port, alc strength by volume >17.9% vol<=18.9% vol, containers<=2 litres 2204212590 Grape wines, nes, incl fort, alc strength by vol >17.9% vol<=18.9% vol,ctnr<=2 L 2204212610 Sherry, alc strength by volume >18.9% vol<=19.9% vol, ctnr<=2 litres 2204212620 Madeira, alc strength by vol >18.9% vol<=19.9% vol, ctnr<=2 litres 2204212630 Port, alc strength by vol >18.9% vol<=19.9% vol, ctnr<=2 litres 2204212690 Grape wines, nes, incl fort, alc strength by vol >18.9% vol<=19.9% vol,ctnr<=2 L 2204212710 Sherry, alc strength by volume >19.9% vol<=20.9% vol, ctnr<=2 litres 2204212720 Port,alc strength by volume >19.9% vol <=20.9% vol,containers <=2 litres 2204212790 Grape wines, nes, incl fort, alc strength by vol >19.9% vol<=20.9% vol,ctnr<=2 L 2204212800 Grape wines, nes, incl fort, alc strength by vol >20.9% vol<=21.9% vol,ctnr<=2 L 2204213110 Sherry,alc strength by volume >21.9% vol<=22.9% vol,containers<=2 litres 2204213120 Madeira,alc strength by volume >21.9% vol<=22.9% vol,containers<=2 litres 2204213130 Port, alc strength by volume >21.9% vol <=22.9% vol, containers <=2 litres 2204213190 Grape wines,nes,incl fort,alc strength by vol >21.9% vol<=22.9% vol,ctnr<=2 L 2204213210 Sherry, alc strength by volume >22.9% vol, containers<=2 litres 2204213220 Madeira, alc strength by volume >22.9% vol, containers<=2 litres 2204213230 Port, alc strength by volume >22.9% vol, containers<=2 litres 2204213290 Grape wines, nes,incl fort, alc strength by vol >22.9% vol, containers<=2 litres 2205101010 Vermouth, white, alc strength by vol <=18.3% vol, ctnr <=2 litres 2205101020 Vermouth, red, alc strength by volume <= 18.3% vol, ctnr <=2 litres 2205101030 Vermouth, rosé, alc strength by volume <=18.3% vol, ctnr <=2 litres 2205101090 Grape wine, nes,flav w plants/arom subs,alc strength by vol<=18.3% vol,ctnr<=2 L 2205102000 Vermouth & flav grape wines,nes,alc strength by vol >18.3%<=22.9%,ctnr<=2 L 2205103000 Vermouth & flav grape wines, nes, alc strength by vol >22.9% vol, ctnr<=2 L 2206002130 Prune wine, fermented, alc strength by vol >7.0% vol<=22.9% vol 2206002200 Prune wine, fermented, alc strength by vol >22.9% vol 2206003100 Perry, sparkling, of an alc strength by vol <= 22.9% vol 2206003900 Perry, sparkling, of an alc strength by vol > 22.9% vol 2206004100 Other sparkling wine, o/t Perry, of an alc strt by vol <= 22.9% vol 2206004900 Other sparkling wine, o/t Perry, of an alc strt by vol > 22.9% vol 221

2206005011 Kosher, blackberry, wine, not sparkling, alc strength by vol<=13.7% vol 2206005019 Fruit wine, nes, not sparkling, alc strength by vol<=13.7% vol 2206005090 Wine, nes, not sparkling, alc strength by vol <=13.7% vol 2206006100 Sake & wine, nes, not sparkling, alc strength by vol >13.7% vol<=14.9% vol 2206006200 Sake & wine, nes, not sparkling, alc strength by vol >14.9% vol<=15.9% vol 2206006300 Sake & wine, nes, not sparkling, alc strength by vol >15.9% vol<=16.9% vol 2206006410 Fruit wine, nes, not sparkling,alc strength by vol >16.9% vol<=17.9% vol 2206006490 Wine, nes, not sparkling, alc strength by vol >16.9% vol<=17.9% vol 2206006500 Sake & wine, nes, not sparkling, alc strength by vol >17.9% vol<=18.9% vol 2206006600 Sake & wine, nes, not sparkling, alc strength by vol >18.9% vol<=19.9% vol 2206006700 Sake & wine, nes, not sparkling, alc strength by vol >19.9% vol<=20.9% vol 2206006800 Sake & wine, nes, not sparkling, alc strength by vol >20.9% vol<=21.9% vol 2206007100 Sake & wine,nes,not sparkling,alc strength by vol >21.9% vol <=22.9% 2206007200 Sake & wine, nes, not sparkling, alc strength by vol >22.9% vol Source: Statistics Canada (CHASS Trade Analyser: Canadian Imports Data. 21 Dec 2008.)

Table A3.13 HS Codes selected to provide the quantity of packaged spirits imported to Ontario, Canada in 2008

Harmonized System Type of packaged spirit imported to Ontario, Canada in 2008 (HS) Code 2208200090 Spirits, obtained by distilling grape wine or grape marc, excluding in bulk 2208300019 Whisky, bourbon, excluding in bulk 2208300029 Whisky, scotch, excluding in bulk 2208300030 Whisky, irish 2208300040 Whisky, rye 2208300099 Whiskies, nes, excluding in bulk 2208401090 Rum, excluding in bulk 2208409000 Other spirits, obtained by distilling fermented sugar cane products 2208500090 Gin and Geneva, excluding in bulk 2208600000 Vodka 2208700000 Liqueurs and cordials 2208901000 Tequila 2208903000 Angostura bitters 2208909200 Fruit brandies Source: Statistics Canada (CHASS Trade Analyser: Canadian Imports Data. 21 Dec 2008.)

Transport distance estimates were obtained through the use of the “World-Ports Distances Calculator” (World Ports Distances Calculator 2010) and Google Maps (Google 2010). Unfortunately, the countries of origin of the wines and spirits may not be the same as the countries where the alcoholic beverage is packaged. Nevertheless, it is assumed unlikely that wines/spirits produced in one country would be shipped to another for packaging and then shipped to Ontario. Model assumptions of specific shipping routes and modes of transportation were generated in consultation with responsible and knowledgeable representatives of the Logistics Department of the LCBO (T. Stewart - LCBO, pers. comm. 2010). Transportation assumptions for wine and spirit imports were as follows: (1) products from Europe, Africa and the Caribbean enter via Port of Montreal and are transported to 222

Toronto via both road and rail (50/50); (2) products from Asia, Oceania and South America travel through the Panama Canal, arrive at the Port of Philadelphia (50%) and the Port of Halifax (50%) and are transported to Toronto via both road and rail (50/50); and (3) products from the United States and Mexico are shipped by road. These assumptions, produced by the author, are the best approximations that could be generated for this component of the life cycle assessment. Table A3.14 lists the overseas wine/spirit import markets, the assumed port of export and port of arrival, the shipping distances between these ports, and the respective volumes of packaged wines and spirits shipped to Toronto in 2008.

Table A3.14 List of overseas wine/spirit import markets, the assumed port of export and port of arrival, the shipping distances between these ports, the respective volumes of packaged wines and spirits shipped to Toronto in 2008

Country of origin Assumed port of Assumed port of Distance Volume of Volume of export arrival (km) packaged wines packaged spirits (litres) (litres) Albania Vlore Montreal 8.26*103 0 395 Antigua and Barbuda San Juan Montreal 4.53*103 2 0 Argentina Buenos Aires 50% Halifax 1.06*104 7.03*105 0 50% Philadelphia 1.09*104 Armenia Batumi Montreal 1.03*104 0 1.11*103 Australia Melbourne 50% Halifax 2.08*104 3.07*106 674 50% Philadelphia 2.00*104 Austria Venice Montreal 8.97*103 1.76*104 3.80*103 Bahamas Freeport Montreal 4.22*103 0 2 Barbados Bridgetown Montreal 5.06*103 0 3.27*104 Belgium Antwerp Montreal 6.10*103 9 9 Bermuda Hamilton Montreal 3.00*103 0 8.71*103 Bolivia Antofagasta 50% Halifax 1.01*104 2 0 50% Philadelphia 9.26*103 Brazil Santos/Rio de 50% Halifax 8.55*103 1.61*103 3 Janeiro 50% Philadelphia 8.87*103 Bulgaria Varna Montreal 9.50*103 1.52*104 1 Cayman Islands Kingston Montreal 4.98*103 0 1 Chile Valparaiso 50% Halifax 1.10*104 1.61*106 0 50% Philadelphia 1.02*104 China Shanghai 50% Halifax 2.20*104 641 0 50% Philadelphia 2.12*104 Columbia Cartagena 50% Halifax 4.03*103 0 861 50% Philadelphia 3.39*103 Costa Rica Limon 50% Halifax 4.38*103 0 0 50% Philadelphia 3.74*103 Croatia Rijeka Montreal 8.87*103 290 8.86*103 Cuba Havana Montreal 4.68*103 0 1.10*105 Cyprus Limassol Montreal 9.39*103 547 588 Czech Republic Hamburg Montreal 6.32*103 0 2.83*103 Denmark Copenhagen Montreal 6.42*103 1.39*103 6.56*103 Dominican Republic Santo Domingo Montreal 4.83*103 0 3.92*103 Estonia Tallinn Montreal 7.40*103 0 1 Finland Kotka Montreal 7.54*103 0 1.04*105 France Le Havre Montreal 5.75*103 2.73*106 5.59*105 Georgia Batumi Montreal 1.03*104 3.49*103 674 223

Germany Hamburg Montreal 6.32*103 3.89*105 3.95*104 Greece Piraeus Montreal 8.63*103 8.71*104 2.49*104 Guatemala Puerto Barrios 50% Halifax 4.07*103 0 2 50% Philadelphia 3.20*103 Guyana Georgetown 50% Halifax 5.50*103 0 1.26*104 50% Philadelphia 4.12*103 Hungary Livorno Montreal 7.49*103 4.85*104 4.06*103 Iceland Reykjavik Montreal 4.52*103 0 1.78*103 India Mumbai 50% Halifax 1.47*104 6 16 50% Philadelphia 1.54*104 Ireland, Republic of Dublin Montreal 5.43*103 0 1.99*105 Israel Tel Aviv Montreal 9.60*103 3.37*104 984 Italy Gioia Tauro Montreal 7.79*103 4.28*106 2.12*105 Jamaica Kingston Montreal 4.98*103 5.44*103 4.10*104 Japan Tokyo 50% Halifax 2.04*104 7.39*104 1 50% Philadelphia 1.96*104 Korea, South Pusan 50% Halifax 2.11*104 7.06*103 362 50% Philadelphia 2.03*104 Latvia Riga Montreal 7.29*103 0 118 Lebanon Beirut Montreal 5.27*103 1.88*103 1.82*103 Lithuania Klaipeda Montreal 7.01*103 0 90 Luxembourg Antwerp Montreal 6.10*103 29 0 Macedonia Bar Montreal 8.40*103 5.64*103 0 (FYROM) Moldova, Republic of Odessa Montreal 9.86*103 2.53*103 2.27*103 Montenegro Trieste Montreal 8.94*103 2.98*104 6.18*103 Morocco Casablanca Montreal 5.85*103 602 0 Netherlands Antwerp Montreal 6.10*103 12 1.62*104 New Zealand Auckland 50% Halifax 1.83*104 2.98*105 1.07*103 50% Philadelphia 1.74*104 Panama Panama City 50% Halifax 3.67*103 0 117 50% Philadelphia 2.80*103 Peru Callao 50% Halifax 8.67*103 0 449 50% Philadelphia 7.80*103 Philippines Manila 50% Halifax 2.03*104 0 1 50% Philadelphia 2.15*104 Poland Gdansk Montreal 6.91*103 4.37*103 9.53*104 Portugal Lisbon Montreal 5.45*103 4.96*105 9.19*103 Romania Constanta Montreal 9.59*103 6.53*103 0 Russian Federation Saint Petersburg Montreal 7.70*103 1.16*103 5.85*105 Saint Kitts Nevis Montreal 4.62*103 0 86 Saint Lucia Bridgetown Montreal 5.06*103 0 5 Slovenia Hamburg Montreal 6.32*103 2.00*103 31 South Africa Cape Town Montreal 1.32*104 6.56*105 2.03*104 Spain La Coruna, Montreal 5.30*103 5.45*105 6.52*103 Sweden Goteborg Montreal 6.22*103 7 3.62*105 (Gothenburg) Switzerland Genoa, Italy Montreal 7.46*103 1.47*103 2.15*104 Thailand Bangkok 50% Halifax 1.94*104 196 0 50% Philadelphia 2.06*104 Trinidad and Tobago Port of Spain Montreal 5.36*103 0 1.05*104 Turkey Izmir Montreal 8.93*103 137 4 Ukraine Odessa Montreal 9.86*103 3.46*103 5.71*103 United Arab Emirates Dubai Montreal 14783 188 708 United Kingdom Felixstowe Montreal 5.96*103 4.64*103 1.10*106 Uruguay Montevideo 50% Halifax 1.04*104 158*103 0 50% Philadelphia 1.07*104 Vietnam Ho Chi Minh 50% Halifax 1.90*104 0 1 City 50% Philadelphia 2.02*104 Virgin Islands, St. Croix Montreal 4.62*103 0 258 British Source of distances: World ports distances. Available at: http://distances.com/. Accessed 2010 02 23.

224

Table A3.15 lists the non-overseas wine/spirit import markets, the assumed location of export, the shipping distances to Toronto, and the respective volumes of packaged wines and spirits shipped to Toronto in 2008.

Table A3.15 List of non-overseas wine/spirit import markets, the land-based shipping distances to the City of Toronto, and the respective volumes of packaged wines and spirits shipped to the city in 2008

Country of origin Assumed location of Distance (km) Volume of packaged Volume of packaged export1 wines (litres) spirits (L) Canada Assumption 2.00*103 4.32*103 8.80*103 (excl. Ontario) Mexico Tijuana (plus 500 km) 4.68*103 1.26*104 1.56*105 Canada (Ontario) Assumption 500 6.72*106 6.94*106 US-Arizona Average 3.49*103 0 0 US-Arkansas Average 1.78*103 2 1.72*103 US-California Napa 4.20*103 2.07*106 1.63*105 US-Connecticut Average 782 7 1.23*103 US-Delaware Average 944 0 0 US-Florida Average 2.18*103 1.16*103 1.26*105 US-Georgia Average 1.80*103 0 1 US-Hawaii Honolulu to Vancouver 4.49*103 Vancouver to Toronto 4.17*103 2 0 US-Idaho Average 3.55*103 3 207 US-Illinois Average 1.04*103 1.67*103 1.63*105 US-Indiana Average 774 0 3 US-Kentucky Average 978 2.04*104 2.02*105 US-Maine Average 972 0 186 US-Maryland Average 923 2.01*103 1.30*104 US-Michigan Average 646 1 1 US-Minnesota Average 1.76*103 11 2.50*104 US-Missouri Average 1.45*103 3 33 US-Nevada Average 3.69*103 0 0 US-New Jersey Average 889 2.07*104 6.72*103 US-New York Average 794 4.40*104 726 US-North Carolina Average 1.30*103 137 0 US-Ohio Average 649 0 1.52*103 US-Oregon Average 3.97*103 7.60*103 0 US-Other States Average 1.84*103 86 11 US-Pennsylvania Average 534 0 0 US-Puerto Rico San Juan 5.07*103 0 3.21*104 US-Tennessee Average 1.35*103 0 1.16*105 US-Texas Average 2.67*103 10 137 US-Virginia Average 1.11*103 332 0 US-U.S. Virgin Is. St. Croix 5.16*103 0 1.65*103 US-Washington, state Average 4.04*103 1.28*104 0 Total 1.73*107 4.65*106 1 “Average” designates the average location in US state, as estimated through Google Maps (first option selected). Available at http://maps.google.ca. Accessed on 2010 02 23. The average distance travelled from Canadian provinces outside of Ontario was assumed to be 2000 km, the approximate distance between Toronto, Ontario and Winnipeg, Manitoba. The average distance travelled from markets within Ontario was assumed to be 150 km, approximately the distance between the City of Toronto and Niagara-on-the-Lake, which is considered a prime destination of Ontario’s “wine country” (Wine Council of Ontario. 2011. Available at: http://winesofontario.org/Niagara-on-the-Lake-en. Accessed on 2011 01 28.).

225

A3.2.3 Transportation of residential waste

Table A3.16 lists the residential waste treatment/disposal/transfer locations and distances from Toronto, mass shipped and tonne-km calculated for shipments to treatment/disposal/transfer locations in 2008.

Table A3.16 Residential waste treatment/disposal locations and estimated distances from Toronto, mass shipped and tonne-km calculated for shipments to treatment/disposal locations in 2008

Type of Treatment / Estimated distance Tonne-km waste disposal / transfer from Toronto 2008 reference Alternative packaging scenario location (km)1 Waste collection All MSW One of seven MSW 20 km one-way trip 2.48*105 1.36*105 collected transfer facilities Landfill Residual Carleton Farms 441 1.60*106 8.99*105 waste Landfill, Michigan Green Lane 203 5.38*104 3.02*104 Landfill, Ontario2 Municipal recycling Glass Belleville, Ontario 130 1.34*105 6.98*104 Brampton, Ontario 43 2.94*104 1.53*104 Stanley, New York 318 1.58*105 8.18*104 Montreal, Quebec 545 2.31*105 1.20*105 Sarnia, Ontario 288 1.36*104 7.05*103 PET Montreal, Quebec 545 1.01*104 1.56*104 Winnipeg, Manitoba 2050 1.14*104 1.75*104 Poly-coat Bannockburn, 887 4.21*102 8.33*102 Illonois Inchon, South 13700 7.91*103 1.56*104 Korea (rail) (rail) 1.67*104 3.29*104 (ship) (ship) Net deposit-return (The Beer Store) recycling Glass Belleville, Ontario 130 7.61*105 3.95*105 Brampton, Ontario 43 1.67*105 8.65*104 Stanley, New York 318 8.92*105 4.63*105 Montreal, Quebec 545 1.31*106 6.80*105 Sarnia, Ontario 288 7.69*104 3.99*104 Mixed Ontario (regional 150 2.28*103 1.53*103 paper distance) China (Shanghai) 13700 1.37*105 9.15*104 (rail) (rail) 3.06*105 2.05*105 (ship) (ship) PET Montreal, Quebec 545 1.28*104 1.97*104 Winnipeg, Manitoba 2050 1.44*104 2.21*104 Poly-coat Bannockburn, 887 8.95*102 1.77*103 Illinois Inchon, South 13700 1.68*104 3.32*104 Korea (rail) (rail) 3.54*104 6.99*104 (ship) (ship) 1 Treatment/disposal/transfer locations supplied by the City of Toronto (2010a), while average distances are estimated through Google Maps (first option selected) (2010) and World ports distances calculator (2010). 2 Approximate location of Green Lane Landfill in St. Thomas, Ontario: 42.813049 degrees latitude, -81.327202 degrees longitude. 226

A3.3 Unit process descriptions

A3.3.1 Selected unit processes

Table A3.17 lists the unit processes selected for the LCA scenarios, the emissions database source, and the year and geography of the technological representation of the processes.

Table A3.17 Unit processes for wine/spirit packaging LCAs

Materials/Processes Unit process description LCA database Year Geography of technology Containers Glass bottles Packaging glass, brown, at Adaptation of Unspecified Default: Avg. European situation plant/55% cullet US-EI based time period -Altered elect. mix (Section A7.3) upon Magaud et al. (2010) Packaging glass, green, at Adaptation of Unspecified Default: Avg. European situation plant/75% cullet US-EI based time period -Altered elect. mix (Section A7.3) upon Magaud et al. (2010) Packaging glass, white, at Adaptation of Unspecified Default: Avg. European situation plant/58% cullet US-EI based time period -Altered elect. mix (Section A7.3) upon Magaud et al. (2010) Aseptic cartons Aluminium, primary, at US-EI Unspecified Avg. European situation plant time period Solid bleached board, SBB, US-EI Unspecified Avg. European situation at plant time period Packaging film, LDPE, at US-EI Unspecified Avg. European situation plant time period Aseptic carton manufacture Author defined 2002 Dijon, France with avg. European unit process electricity PET bottles Polyethylene terephthalate, US-EI Unspecified Default: Avg. European situation granulate, bottle grade, at time period with US electricity plant -Altered elect. mix (Section A7.3) Injection moulding US-EI Unspecified Default: Avg. European situation (Preforms) time period with US electricity -Altered elect. mix (Section A7.3) Blow moulding US-EI Unspecified Default: Avg. European situation time period with US electricity -Altered elect. mix (Section A7.3) Secondary packaging (capsules, closures and paper labels) Capsules Aluminium, primary, at US-EI Unspecified Avg. European situation with US plant time period electricity Polyethylene terephthalate, US-EI Unspecified Avg. European situation with US granulate, amorphous, at time period electricity plant Polyvinylchloride, at US-EI Unspecified Avg. European situation with US regional storage (transport time period electricity emissions removed) Tin, at regional storage US-EI Unspecified Avg. European situation with US time period electricity Closures Aluminium, primary, at US-EI Unspecified Avg. European situation with US plant time period electricity Kraft paper, bleached, at US-EI 2000 Avg. European situation with US plant electricity Polyethylene, HDPE, US-EI Unspecified Avg. European situation with US granulate, at plant time period electricity 227

Polyethylene, LDPE, US-EI Unspecified Avg. European situation with US granulate, at plant time period electricity Polyethylene terephthalate, US-EI Unspecified Avg. European situation with US granulate, amorphous, at time period electricity plant Polypropylene, granulate, at US-EI Unspecified Avg. European situation with US plant time period electricity Polyvinylchloride, at US-EI Unspecified Avg. European situation with US regional storage (transport time period electricity emissions removed) Raw cork, at forest road EcoInvent 1993 Avg. European situation Tin, at regional storage US-EI Unspecified Avg. European situation with US time period electricity Paper labels for glass Paper, wood-containing, US-EI Unspecified Avg. European situation with US and PET wine and LWC, at regional storage time period electricity spirits containers Transportation Transportation by Transport, lorry >28t, fleet US-EI Unspecified Switzerland with US electricity truck average time period Transportation by rail Transport, freight, rail, US-EI Unspecified United States diesel time period Transportation by Transport, transoceanic US-EI Unspecified Global (extrapolated from a port in ship freight ship time period Netherlands) Waste management Residential waste Transport, municipal waste US-EI Unspecified Switzerland with US electricity collection collection, lorry 21t time period Recyclable material Recyclables, sorted at MRF, Author defined Unspecified Avg. European situation with sorting at MRF for further treatment unit process time period Ontario electricity supply mix (excludes transport) based upon from 2008 (see Section A7.3) US-EI (original process: waste paper, sorted, for further treatment/RER with US electricity U) Residential waste Transport, freight, rail, US-EI Unspecified United States transportation diesel time period Transport, lorry >28t, fleet US-EI Unspecified Switzerland with US electricity average time period Transport, transoceanic US-EI Unspecified OCE (Oceanic) freight ship time period Landfilling Landfill US-EI Unspecified Switzerland with US electricity time period (for US landfill) -Ontario electricity mix for Ontario landfill Bottle washing Bottle washing Author defined 2010 Global (extrapolated from data unit process supplied by a wine bottle washing equipment manufacturer)

A3.3.2 Author-defined unit processes

Note: The following author-defined unit process is described in Appendix 4 (Section A4.4.2: recyclables, sorted at MRF, for further treatment/Ontario 2008.

228

A3.3.2.1 Aseptic carton manufacture

The most recent LCA data on wine aseptic carton manufacture was published by Franklin Associates (2006), which specifies energy inputs of approximately 0.17 million BTU per 1000 one litre tetra prisma containers. However, the water inputs and environmental emissions from this process are not listed, which limits the utility of the supplied data for define an LCA unit process. Thus, an older dataset on generic aseptic carton manufacture that was published in 2003 by Tetra Pak Inc. is used (Tetra Pak 2003). The processes associated with the production of the inks for the containers were excluded from the aseptic carton manufacturing process because the chemical composition of the inks is not known.

Source of data: 2002 data Tetra Pak France (Dijon Tetra Pak Manufacturing Facility) Electricity use: 9.8 kwh/1000 packages Water consumption: 6.18 litres of water/1000 packages Ink consumption (excluded from process): 0.08 g of ink per standard package Loss at manufacturing facility: 5.2%, with 99.4% recycled and 0.06% incinerated Wastewater output: 709mg/litre DOC in wastewater

Specifications of defined process 1: Known outputs to technosphere: 1 litre aseptic carton manufacturing; 36.4 kg (1000 containers with 5.2% loss) Known inputs from nature (resources): Water, process, unspecified natural origin; 0.00618 m³ Known inputs from technosphere (electricity/heat): Electricity, medium voltage, at grid/DE U; 9.8 kwh Known outputs to technosphere. Waste and emissions to treatment: Treatment, sewage, unpolluted, to wastewater treatment, class 3/CH with US electricity U; 6.18 litres

Specifications of defined process 2: Known outputs to technosphere: 1.5 litre aseptic carton manufacturing; 52.7 kg (1000 containers with 5.2% loss) Known inputs from nature (resources): Water, process, unspecified natural origin; 0.00618 m³ Known inputs from technosphere (electricity/heat): Electricity, medium voltage, at grid/DE U; 9.8 kwh Known outputs to technosphere. Waste and emissions to treatment: Treatment, sewage, unpolluted, to wastewater treatment, class 3/CH with US electricity U; 6.18 litres 229

A3.3.2.2 Bottle washing

Within the EcoInvent database, as well as the others supplied with the SimaPro 7.2 LCA software, there was only one unit process available that depicted bottle washing. Unfortunately, this unit process, supplied through the BUWAL 250 database, represented technology from the early 1990s. Due to the age of the available data, bottle washing data from modern equipment were acquired from a bottle washing equipment manufacturer (name withheld due to a confidentiality agreement), which revealed very significant improvements in efficiency. These data were used in the design of the bottle washing unit process used for the alternative packaging scenario of the case study. Due to lack of sufficient information, the environmental burdens caused by the production of the bottle washing machine and facility were excluded from the defined process.

Source of data: A bottle washing equipment manufacturer (permission to use data was acquired with the stipulation that the source remains confidential)

Technical description of equipment -Nominal capacity of 30 000 bottles/hr (1 litre bottles) -Electricity consumption: 60 kWh/30000 bottles -Heat consumption with insulation: 1128 MJ/h (assumption of natural gas used for heating source) -Water consumption: 190-220 ml/bottle -Caustic concentration: 2% (author’s estimate of 3.8 g NaOH per bottle) -Machine weight: 52.7 tonnes (material composition unknown – excluded from bottle washing unit process)

Specifications of defined process: Known outputs to technosphere: Bottle washing (1 washed bottle) Known inputs from nature (resources): Water, unspecified natural origin; 0.19 kg Known inputs from technosphere (materials/fuels): NaOH(100%); 0.0038 kg Known inputs from technosphere (electricity/heat): Electricity, medium voltage, at grid/Ontario 2008; 0.002 kwh Heat, natural gas, at boiler atm. low-NOx condensing non-modulating <100kw/RER with US electricity U; 0.0376 MJ Known outputs to technosphere. Waste and emissions to treatment: 230

Treatment, sewage, unpolluted, to wastewater treatment, class 3/CH with US electricity U; 0.00019 m³

A3.3.2.3 Glass production and recycling

In their LCA of beer containers in Quebec, Canada, Magaud et al. (2010) adapted the EcoInvent unit processes for the production of brown, green and “white” (clear) glass so that they would represent conditions using “electric boost” furnaces, which are the majority of glass furnaces. Also provided was a method to adapt the original unit processes to take varying levels of glass cullet inputs into account. The unit processes designed for glass are based on these adaptations, which affected energy inputs (electricity and natural gas) as well as inputs of silica sand, limestone, soda, dolomite and feldspar. In order to represent glass recycling, the method by Magaud et al. (2010) was also used to design a figurative glass production process using 100% cullet.

A3.3.2.4 PET recycling

Due to the cut-off allocation adopted in the US-EI database, and the lack of a recent unit process for PET recycling, data from a published LCA of plastics recycling (Franklin Associates 2010) were used to formulate an LCA process for the recycling of PET. The reprocessing efficiency (80%) of PET recycling in this LCA was adopted. Environmental burdens associated with surfactants, defoamers, and wetting agents needed to be omitted because their chemical compositions were not defined in the LCA, and representative unit processes were not available in the LCA databases. Environmental burdens caused by the production of the recycling equipment and facility are excluded from the defined process.

Source of data: Franklin Associates 2010

Specifications of defined process: Known outputs to technosphere: Recycling of PET; 1000 lb Known inputs from nature (resources): 231

Water, process and cooling, unspecified natural origin, 47.3 gal Known inputs from technosphere (materials/fuels): Sodium hydroxide, production mix, at plant/kg NREL/RNA; 23.8 lb Electricity, medium voltage, at grid/Ontario 2008; 208 kwh Natural gas in industrial equipment (Franklin Associates 1998); 1207 cuft LPG into industrial boilers; 0.031 gal Propane/butane, at refinery/RER with US electricity U; 0.0101 kg Emissions to air: Particulates; 0.039 lb VOC, volatile organic compounds; 0.037 lb Known outputs to technosphere. Waste and emissions to treatment: Treatment, sewage, unpolluted, to wastewater treatment, class 3/CH with US electricity U; 47.3 gal Emissions to water (excluded from process, listed for information purposes): BOD5, Biological Oxygen Demand; 7.26 lb COD, Chemical Oxygen Demand; 20.2 lb Suspended solids, unspecified; 2.98 lb

A3.3.3 Electricity production mixes

In order to improve the level of representation of the WasteMAP scenario results, the electricity generation mixes (i.e., the mix of fuel sources used to generate the electricity supplied) of the countries or province in which those processes within the LCA system boundary are undertaken are substituted for the default mixes in the database processes. For those downstream processes taking place in Ontario, the 2008 Ontario electricity generation mix is substituted for the US mix, as defined in the US-EI database. This substitution only applies to first order processes. Should the production of the inputs (e.g., capital equipment) to the downstream processes also require electricity, the default US electricity mix is assumed. Transportation processes also use the default US- EI assumptions. The upstream unit processes for the production of the glass and plastic wine and spirit packaging used the mean 2008 electricity mix of the countries of origin of the wines and spirits, except for domestic products, which uses the Ontario 2008 mix. The only exceptions are those processes associated with the materials for aseptic cartons and corks, which use the average European electricity mixes. Closures (excluding corks), capsules and labels use the US electricity mix from the US-EI database. 232

Electricity supply mix data from Ontario and all countries supplying 1% or more of the imports of wines and spirits are selected to estimate the mean electricity supply mix used in the production of wine/spirit packaging. Due to a lack of data on the origins of the containers used to package the wines/spirits from each country, it is assumed that the packaging was manufactured using the electricity supply mix of the country supplying the wine/spirit. The procedure used to generate these estimates is as follows: (1) the electricity supply mixes of the countries that supply 1% or more of wine and spirit volume imports in 2008 were obtained from International Energy Agency statistics from 2008 (see Table A3.18); (2) for each selected country, the percent representation of the sample of countries in terms of the volumes of wine and spirit imports supplied was calculated; (3) for each respective country, these percent representations were multiplied by the percentage of the electricity supplied by each type of electricity generation; (4) the results from (3) for each country were summed in order to estimate the percent composition of the overall electricity mix used to produce the containers for imported packaged wines and imported packaged spirits; and (5) for each scenario, these electricity mix results for wine and spirit import markets were combined by multiplying them by the proportions of the primary packaging mass used for wines and for spirits, and then summing them (see Table 3.19 supply mix results). For domestic packaged wines and spirits, the 2008 electricity supply mix of the Ontario power grid was obtained from Ontario’s Independent Electricity System Operator (2009).

Table A3.18 Percentage composition of the electricity production mixes of Ontario and wine and spirit import markets

Country % of 2008 electricity production mix1 Nuclear Coal Gas Hydro Oil Wind Biomass Waste Solar Other PV Argentina 6.01 2.30 53.40 25.22 11.70 0.03 1.34 0.00 0.00 0.00 Australia 0.00 76.82 14.97 4.69 1.07 1.53 0.86 0.00 0.06 0.00 Chile 0.00 23.64 3.66 40.52 26.95 0.06 5.16 0.00 0.00 0.00 Cuba 0.00 0.00 0.00 0.78 97.01 0.00 2.21 0.00 0.00 0.00 Finland 29.65 18.48 14.52 22.10 0.55 0.34 13.10 0.61 0.01 0.65 France 76.45 4.74 3.81 11.89 1.01 0.99 0.37 0.66 0.01 0.09 Germany 23.30 45.61 13.76 4.23 1.45 6.37 3.12 1.47 0.69 0.00 Ireland 0.00 27.01 54.12 4.38 5.83 8.12 0.54 0.00 0.00 0.00 Italy 0.00 15.23 54.12 14.80 9.86 1.52 1.38 1.02 0.06 2.02 Mexico 3.79 8.27 50.63 15.13 19.05 0.10 0.31 0.00 0.00 2.73 New 0.00 11.01 24.33 50.97 0.30 2.41 1.27 0.00 0.00 9.71 Zealand Ontario, 53.0 14.5 6.92 24.1 - 0.9 - - - 0.6 Canada (gas/oil) 233

(2008) Poland 0.00 91.80 2.03 1.76 1.49 0.54 2.21 0.18 0.00 0.00 Portugal 0.00 24.36 33.06 15.87 9.02 12.52 3.42 1.24 0.08 0.42 Russia 15.68 18.91 47.55 16.02 1.55 0.00 0.00 0.24 0.00 0.04 South 5.03 93.23 0.00 1.56 0.06 0.01 0.10 0.00 0.01 0.00 Africa Spain 18.80 15.93 38.75 8.32 5.74 10.26 0.79 0.50 0.82 0.10 Sweden 42.58 1.49 0.40 46.13 0.58 1.33 6.04 1.44 0.00 0.00 United 13.48 32.54 45.39 2.37 1.57 1.82 2.08 0.74 0.00 0.00 Kingdom United 19.18 48.81 20.84 6.45 1.32 1.27 1.15 0.51 0.04 0.43 States United 19.74 49.68 17.42 8.21 3.34 0.35 1.14 Included 0.02 0.10 States within (2004)3 “Other” 1 Sources: Independent Electricity System Operator (2009); International Energy Agency (2010). 2 Data available on the Ontario electricity supply mix combines natural gas and oil sources into one category. Within the LCA, this component of the mix is identified as natural gas. 3 This electricity mix is used as the default for all unit processes within the US-EI database.

Table A3.19 US-EI unit processes for electricity production categories, and the average electricity production mix for the upstream component of each LCA scenario

IEA electricity Unit processes % composition of 2008 electricity mix for wine/spirit packaging systems production (US-EI database) 2008 Reference Alternative Scenario source categories Import markets Domestic market Import markets Domestic market Nuclear Electricity, 31.7 53.0 30.6 53.0 nuclear, at power plant/US with US electricity U Coal Electricity, hard 24.9 14.5 25.5 14.5 coal, at power plant/US with US electricity U Gas Electricity, 18.7 6.9 17.1 6.9 natural gas, at power plant/US with US electricity U Hydro Electricity, 17.6 24.1 17.4 24.1 hydropower, at power plant/CH with US electricity U Oil Electricity, oil, at 3.8 - 4.0 - power plant/UCTE with US electricity U Wind Electricity, at 1.4 0.9 1.5 0.9 wind power plant/RER U Biomass Electricity, at 1.0 - 1.0 - cogen 6400kWth, wood, allocation exergy/CH with US electricity U Waste Electricity, from 0.3 - 0.3 - waste, at municipal waste incineration plant/CH with US electricity U 234

Solar PV Electricity, 0 - 0 - production mix photovoltaic, at plant/US U Geothermal Not available 0.4 - 0.4 - Solar thermal Not available 0 - 0 - Tide Not available 0 - 0 - Other sources Not available 0.3 0.6 0.3 0.6 Total, excluding undefined unit 99.4% 100% 99.3% ‘ processes1 1 Note: Total may not add up due to rounding

A3.3.4 Recycling and avoided burdens

Table A3.20 Unit processes for the recycling component of the wine/spirit packaging LCA

Materials Unit process description LCA database Year Geography of Technology Aluminum Avoided product US-EI Unspecified Avg. European situation Primary aluminium time period with US electricity Old aluminium scrap US-EI Unspecified Avg. European situation time period with US electricity Cardboard Avoided product US-EI Unspecified Avg. European situation Core board time period with US electricity Corrugated board, recycling US-EI Unspecified Avg. European situation fibre (single wall) time period with US electricity Glass Avoided product: Adaptation of US-EI Unspecified Avg. European situation Packaging glass, brown, at based upon Magaud et time period -Altered elect. mix plant/Virgin al. (2010) (Section A7.3) 235

Packaging glass, brown, at Adaptation of US-EI Unspecified Avg. European situation plant/100% Cullet based upon Magaud et time period -Altered elect. mix al. (2010) (Section A7.3) Avoided product: Adaptation of US-EI Unspecified Avg. European situation Packaging glass, green, at based upon Magaud et time period -Altered elect. mix plant/Virgin al. (2010) (Section A7.3) Packaging glass, green, at Adaptation of US-EI Unspecified Avg. European situation plant/100% Cullet based upon Magaud et time period -Altered elect. mix al. (2010) (Section A7.3) Avoided product: Adaptation of US-EI Unspecified Avg. European situation Packaging glass, white, at based upon Magaud et time period -Altered elect. mix plant/Virgin al. (2010) (Section A7.3) Packaging glass, white, at Adaptation of US-EI Unspecified Avg. European situation plant/100% Cullet based upon Magaud et time period -Altered elect. mix al. (2010) (Section A7.3) Paper Avoided product US-EI Unspecified Avg. European situation Paper, newsprint, 0% DIP, time period with US electricity at plant Paper, newsprint, at plant US-EI Unspecified Avg. European situation time period with US electricity PET Avoided product US-EI Unspecified Avg. European situation Polyethylene terephthalate, time period with US electricity granulate, bottle grade, at plant PET recycling Author-defined unit 2010 United States process based upon Franklin Associates 2010

A3.4 Unit process inputs

A3.4.1 Individual package scale

For each scenario at the scale of the individual package, Tables A3.21, A3.22, A3.23 and A3.24 list the LCA process inputs to calculation net environmental burdens, and specify the quantities of each unit process required.

Table A3.21 Upstream unit processes and the quantities required under each one litre wine packaging LCA scenario

Process Name of process Unit AC CSU LSU PET RFG category Containers Aseptic cartons Aluminium, primary, Gram 2.15 0 0 0 0 at plant Solid bleached board, Gram 26.90 0 0 0 0 SBB, at plant Packaging film, Gram 8.09 0 0 0 0 LDPE, at plant 1 litre aseptic carton Gram 37.14 0 0 0 0 manufacture

Glass Packaging glass, Gram 0 6.65 5.32 0 0.44 brown, at plant/55% 236

Cullet Packaging glass, Gram 0 383.98 307.18 0 25.60 green, at plant/75% Cullet Packaging glass, Gram 0 163.45 130.76 0 10.90 white, at plant/58% Cullet PET Polyethylene Gram 0 0 0 59.18 0 terephthalate, granulate, bottle grade, at plant Blow molding Gram 0 0 0 59.18 0 Injection molding Gram 0 0 0 59.18 0 Secondary packaging (capsules, closures and paper labels) Capsules Aluminium, primary, Gram 0 0.14 0.14 0 0.14 at plant Packaging film, Gram 0 0.11 0.11 0 0.11 LDPE, at plant Polyvinylidenchloride Gram 0 0.02 0.02 0 0.02 (PVDC), granulate, at plant Tin, at regional Gram 0 0.05 0.05 0 0.05 storage Closures Aluminium, primary, Gram 0 3.34 3.34 4.14 3.34 at plant Kraft paper, bleached, Gram 0 0.02 0.02 0.02 0.02 at plant Polyethylene Gram 0 0.26 0.26 0.32 0.26 terephthalate, granulate, amorphous, at plant Polyethylene, HDPE, Gram 1.1 0.09 0.09 0 0.09 granulate, at plant Polyethylene, LDPE, Gram 0.16 0.16 0 0.16 granulate, at plant Polypropylene, Gram 1.1 0.09 0.09 0 0.09 granulate, at plant Polyvinylidenchloride Gram 0 0.02 0.02 0.03 0.02 (PVDC), granulate, at plant Raw cork, at forest Gram 0 0.31 0.31 0 0.31 road Steel, cold rolled Gram 0 0.14 0.14 0 0.14 sheet, at plant NREL/RNA Tin, at regional Gram 0 0.07 0.07 0 0.07 storage Paper labels Paper, wood- Gram 0 1.9 1.9 1.9 1.9 containing, LWC, at regional storage Upstream transportation Container Transport, freight, Tonne- 3.56*10-2 5.33*10-1 4.27*10-1 6.25*10-2 4.16*10-2 transportation rail, diesel km from Transport, lorry >28t, Tonne- 4.20*10-2 6.29*10-1 5.05*10-1 7.38*10-2 4.91*10-2 manufacturer to fleet average km winery/distillery Transport, Tonne- 3.71*10-1 5.56*100 4.46*100 6.52*10-1 4.34*10-1 and Toronto transoceanic freight km product ship distribution centre

237

Table A3.22 Upstream unit processes and the quantities required under each 750 ml spirit packaging LCA scenario

Process Name of process Unit CSU LSU PET RFG category Containers Glass Packaging glass, Gram 92.49 73.99 0 6.17 brown, at plant/55% Cullet Packaging glass, Gram 23.12 18.50 0 1.54 green, at plant/75% Cullet Packaging glass, Gram 409.90 327.92 0 27.33 white, at plant/58% Cullet PET Polyethylene Gram 0 0 64.29 0 terephthalate, granulate, bottle grade, at plant Blow molding Gram 0 0 64.29 0 Injection molding Gram 0 0 64.29 0 Secondary packaging (closures and paper labels) Closures Aluminium, primary, Gram 1.57 1.57 0 1.57 at plant Kraft paper, bleached, Gram 0.01 0.01 0 0.01 at plant Polyethylene Gram 0.12 0.12 0 0.12 terephthalate, granulate, amorphous, at plant Polyethylene, HDPE, Gram 0.55 0.55 1.33 0.55 granulate, at plant Polyethylene, LDPE, Gram 0.01 0.01 0 0.01 granulate, at plant Polypropylene, Gram 0.55 0.55 1.33 0.55 granulate, at plant Polyvinylidenchloride Gram 0.01 0.01 0 0.01 (PVDC), granulate, at plant Raw cork, at forest Gram 0.33 0.33 0 0.33 road Steel, cold rolled Gram 0.73 0.73 0 0.73 sheet, at plant NREL/RNA Tin, at regional Gram 0.04 0.04 0 0.04 storage Paper labels Paper, wood- Gram 1.9 1.9 1.9 1.9 containing, LWC, at regional storage Upstream transportation Container Transport, freight, Tonne- 5.15*10-1 4.13*10-1 6.67*10-2 3.96*10-2 transportation rail, diesel km from Transport, lorry >28t, Tonne- 6.08*10-1 4.88*10-1 7.88*10-2 4.67*10-2 manufacturer to fleet average km winery/distillery Transport, Tonne- 5.37*100 4.31*100 6.96*10-1 4.13*10-1 and Toronto transoceanic freight km product ship distribution centre

238

Table A3.23 Downstream unit processes and the quantities required under each one litre wine packaging LCA scenario

Process Name of process Unit AC CSU LSU PET RFG category Downstream transportation Waste Transport, Tonne-km 4.71*10-4 2.11*10-3 1.71*10-3 5.63*10-4 2.30*10-4 collection municipal waste collection, lorry 21t Landfill Transport, lorry Tonne-km 1.01*10-2 4.54*10-2 3.68*10-2 1.21*10-2 4.94*10-3 >28t, fleet average Recycling Transport, Tonne-km 1.09*10-4 4.52*10-3 3.62*10-3 3.67*10-4 3.19*10-4 freight, rail, diesel Transport, lorry Tonne-km 2.58*10-3 1.07*10-1 8.57*10-2 8.70*10-3 7.56*10-3 >28t, fleet average Transport, Tonne-km 2.41*10-4 1.00*10-2 8.02*10-3 8.14*10-4 7.08*10-4 transoceanic freight ship Round trip: Transport, lorry Tonne-km 0 0 0 0 1.20*10-1 Used >28t, fleet containers average between Toronto and refilling facility Waste treatment Landfilling Landfill/CH with Gram 23.84 106.98 86.67 28.53 11.63 US electricity Recycling Sorting Recyclables, Gram 4.11 80.43 64.35 18.80 5.36 sorted at MRF, for further treatment/Ontario 2008 Glass Avoided product: Gram 0 4.50 3.60 0 0.30 Packaging glass, brown, at plant/Virgin Packaging glass, Gram 0 4.50 3.60 0 0.30 brown, at plant/100% Cullet Avoided product: Gram 0 326.41 261.12 0 21.76 Packaging glass, green, at plant/Virgin Packaging glass, Gram 0 326.41 261.12 0 21.76 green, at plant/100% Cullet Avoided product: Gram 0 110.53 88.42 0 7.37 Packaging glass, white, at plant/Virgin Packaging glass, Gram 0 110.53 88.42 0 7.37 white, at plant/100% Cullet Paper Avoided product Gram 0 1.49 1.49 0.73 2.12 239

Paper, newsprint, 0% DIP, at plant Paper, newsprint, Gram 0 1.49 1.49 0.73 2.12 at plant Polycoat Avoided product: Gram 10.98 0 0 0 0 Core board Corrugated Gram 10.98 0 0 0 0 board, recycling fibre (single wall) PET Avoided product: Gram 0 0 0 28.26 0 Polyethylene terephthalate, granulate, bottle grade Process based Gram 0 0 0 28.26 0 upon Franklin Associates 2010 Bottle washing Bottle Bottle washing Bottles 0 0 0 0 0.93 washing (author defined)

Table A3.24 Downstream unit processes and the quantities required under each 750 ml spirit packaging LCA scenario

Process Name of process Unit CSU LSU PET RFG category Downstream transportation Waste Transport, Tonne-km 1.99*10-3 1.61*10-3 5.64*10-4 2.16*10-4 collection municipal waste collection, lorry 21t Landfill Transport, lorry Tonne-km 4.28*10-2 3.47*10-2 1.21*10-2 4.64*10-3 >28t, fleet average Recycling Transport, Tonne-km 4.29*10-3 3.43*10-3 3.98*10-4 2.98*10-4 freight, rail, diesel Transport, lorry Tonne-km 1.02*10-1 8.13*10-2 9.43*10-3 7.06*10-3 >28t, fleet average Transport, Tonne-km 9.50*10-3 7.61*10-3 8.83*10-4 6.61*10-4 transoceanic freight ship Round trip: Transport, lorry Tonne-km 0 0 0 1.12*10-1 Used >28t, fleet containers average between Toronto and refilling facility Waste treatment Landfilling Landfill/CH with Gram 100.83 81.57 28.54 10.92 US electricity Recycling Sorting Recyclables, Gram 76.29 61.03 20.42 5.09 sorted at MRF, for further treatment/Ontario 2008 Glass Avoided product: Gram -62.54 -50.03 0 -4.17 Packaging glass, brown, at 240

plant/Virgin Packaging glass, Gram 62.54 50.03 0 4.17 brown, at plant/100% Cullet Avoided product: Gram -78.95 -63.16 0 -5.26 Packaging glass, green, at plant/Virgin Packaging glass, Gram 78.95 63.16 0 5.26 green, at plant/100% Cullet Avoided product: Gram -277.17 -221.74 0 -18.48 Packaging glass, white, at plant/Virgin Packaging glass, Gram 277.17 221.74 0 18.48 white, at plant/100% Cullet Paper Avoided product Gram -1.49 -1.49 -0.73 -1.49 Paper, newsprint, 0% DIP, at plant Paper, newsprint, Gram 1.49 1.49 0.73 1.49 at plant PET Avoided product: Gram 0 0 -30.70 0 Polyethylene terephthalate, granulate, bottle grade Process based Gram 0 0 30.70 0 upon Franklin Associates 2010 Bottle washing Bottle washing Bottle washing Bottles 0 0 0 0.93 (author defined)

A3.4.2 Municipal scale

For each municipal scale scenario, Tables A3.25 and A3.26 list the LCA processes used in the calculation of net environmental burdens, and specify the quantities of each unit process required.

Table A3.25 Upstream unit processes and the quantities required under each municipal scale wine/spirit packaging LCA scenario

Process category Name of process Unit 2008 reference Alternative scenario packaging scenario Containers Aseptic cartons Aluminium, primary, Tonne 1.03 1.79 at plant Solid bleached board, Tonne 12.94 26.97 SBB, at plant Packaging film, Tonne 3.86 6.47 LDPE, at plant 241

1 litre aseptic carton Tonne of aseptic 17.54 17.91 manufacture carton inputs 1.5 litre aseptic carton Tonne of aseptic 0.29 17.32 manufacture carton inputs Glass bottles Packaging glass, Tonne 1354.11 564.92 brown, at plant/55% Cullet Packaging glass, Tonne 9267.89 5358.48 green, at plant/75% Cullet Packaging glass, Tonne 9130.58 4332.05 white, at plant/58% Cullet PET bottles Polyethylene Tonne 91.38 140.71 terephthalate, granulate, bottle grade, at plant Blow molding Tonne of PET inputs 91.38 140.71 Injection molding Tonne of PET inputs 91.38 140.71 Secondary packaging (capsules, closures and paper labels) Capsules Aluminium, primary, Tonne 7.08 6.73 at plant Packaging film, Tonne 6.26 5.95 LDPE, at plant Polyvinylidenchloride Tonne 2.22 2.11 (PVDC), granulate, at plant Tin, at regional Tonne 5.62 5.13 storage Closures Aluminium, primary, Tonne 60.78 59.04 at plant Kraft paper, bleached, Tonne 0.34 0.33 at plant Polyethylene Tonne 4.73 4.60 terephthalate, granulate, amorphous, at plant Polyethylene, HDPE, Tonne 14.95 15.03 granulate, at plant Polyethylene, LDPE, Tonne 22.07 21.90 granulate, at plant Polypropylene, Tonne 14.95 15.03 granulate, at plant Polyvinylidenchloride Tonne 0.41 0.39 (PVDC), granulate, at plant Raw cork, at forest Tonne 39.30 38.89 road Steel, cold rolled Tonne 10.25 9.19 sheet, at plant NREL/RNA Tin, at regional Tonne 1.35 1.31 storage Labels – paper (for Paper, wood- Tonne 70.43 69.81 glass and PET wine containing, LWC, at and spirits containers) regional storage Upstream transportation Container Transport, freight, rail, Tonne-km 8.29*106 4.41*106 transportation from diesel manufacturer to Transport, lorry >28t, Tonne-km 8.29*106 4.41*106 winery/distillery fleet average Transport, Tonne-km 1.23*107 6.53*106 transoceanic freight ship 242

Container Transport, freight, rail, Tonne-km 1.10*107 6.33*106 transportation from diesel winery/distillery to Transport, lorry >28t, Tonne-km 1.44*107 8.22*106 Toronto product fleet average distribution centre Transport, Tonne-km 1.87*108 1.04*108 transoceanic freight ship

Table A3.26 Downstream unit processes and the quantities required under each municipal scale wine/spirit packaging LCA scenario

Process category Name of process Unit 2008 reference Alternative scenario packaging scenario Downstream transportation Waste collection Transport, municipal Tonne-km 2.48*105 1.36*105 waste collection, lorry 21t Landfill Transport, lorry >28t, Tonne-km 1.66*106 9.29*105 fleet average Recycling Transport, freight, rail, Tonne-km 1.62*105 1.40*105 diesel Transport, lorry >28t, Tonne-km 3.83*106 2.04*106 fleet average Transport, Tonne-km 3.58*105 3.07*105 transoceanic freight ship Round trip: Used Transport, lorry >28t, Tonne-km 0 1.24*106 containers between fleet average Toronto and refilling facility Waste treatment Landfilling Landfill/CH with US Tonne 3.90*103 2.19*103 electricity Recycling Sorting Recyclables, sorted at Tonne 2.87*103 1.53*103 MRF, for further treatment/Ontario 2008 Glass Avoided product: Tonne -9.16*102 -3.82*102 Packaging glass, brown, at plant/Virgin Packaging glass, Tonne 9.16*102 3.82*102 brown, at plant/100% Cullet Avoided product: Tonne -8.63*103 -4.85*103 Packaging glass, green, at plant/Virgin Packaging glass, Tonne 8.63*103 4.85*103 green, at plant/100% Cullet Avoided product: Tonne 6.17*103 2.93*103 Packaging glass, white, at plant/Virgin Packaging glass, Tonne 6.17*103 2.93*103 white, at plant/100% Cullet Paper Avoided product Tonne -5.40*101 -3.61*101 Paper, newsprint, 0% DIP, at plant Paper, newsprint, at Tonne 5.40*101 3.61*101 plant Polycoat Avoided product: Tonne -5.55*100 -1.05*101 Core board 243

Corrugated board, Tonne 5.55*100 1.05*101 recycling fibre (single wall) PET Avoided product: Tonne -4.04*101 -6.21*101 Polyethylene terephthalate, granulate, bottle grade Process based upon Tonne 4.04*101 6.21*101 Franklin Associates 2010 Bottle washing Bottle washing Bottle washing Bottles 0 1.11*107 (author defined)

A3.5 Midpoint level impacts

A3.5.1 Individual package scale

Table A3.27 Midpoint level environmental impacts from the individual wine packaging scenarios under the ReCiPe (H) LCIA method.

Impact category Unit AC CSU LSU PET RFG -1 -1 -1 -1 -1 Climate change kg CO2 eq 1.49*10 7.85*10 6.68*10 3.52*10 1.32*10 Ozone depletion kg CFC-11 5.94*10-9 6.17*10-8 5.39*10-8 6.56*10-8 1.15*10-8 eq -4 -3 -3 -3 -4 Terrestrial kg SO2 eq 4.49*10 5.38*10 4.69*10 1.87*10 8.22*10 acidification Freshwater kg P eq 4.29*10-6 2.83*10-5 2.44*10-5 1.53*10-5 7.88*10-6 eutrophication Marine kg N eq 1.59*10-4 4.86*10-4 4.32*10-4 2.60*10-4 8.03*10-5 eutrophication Human toxicity kg 1,4-DB 1.61*10-2 9.30*10-2 7.76*10-2 3.98*10-2 1.59*10-2 eq Photochemical kg NMVOC 4.61*10-4 4.10*10-3 3.66*10-3 1.24*10-3 6.84*10-4 oxidant formation -4 -3 -3 -4 -4 Particulate matter kg PM10 eq 1.87*10 1.82*10 1.60*10 6.07*10 3.22*10 formation Terrestrial kg 1,4-DB 9.87*10-6 1.48*10-4 1.23*10-4 3.20*10-5 1.81*10-5 ecotoxicity eq Freshwater kg 1,4-DB 8.87*10-4 1.17*10-3 1.04*10-3 1.56*10-3 4.01*10-4 ecotoxicity eq Marine ecotoxicity kg 1,4-DB 8.76*10-4 1.41*10-3 1.26*10-3 1.31*10-3 4.34*10-4 eq Ionising radiation kg U235 eq 2.10*10-2 5.66*10-2 4.90*10-2 9.94*10-2 1.49*10-2 Agricultural land m²a 1.27*10-1 1.42*10-1 1.15*10-1 5.93*10-2 1.30*10-2 occupation Urban land m²a 2.39*10-3 5.62*10-3 4.91*10-3 3.15*10-3 1.07*10-3 occupation Natural land m² 3.11*10-5 1.65*10-4 1.50*10-4 6.35*10-5 4.06*10-5 transformation Water depletion m³ 2.52*10-3 1.39*10-3 1.28*10-3 1.81*10-3 7.07*10-4 Metal depletion kg Fe eq 3.16*10-3 2.04*10-1 2.01*10-1 1.53*10-1 1.85*10-1 Fossil depletion kg oil eq 3.80*10-2 1.97*10-1 1.71*10-1 1.33*10-1 3.73*10-2

244

Table A3.28 Midpoint level environmental impacts from the individual spirit packaging scenarios under the ReCiPe (H) LCIA method.

Impact category Unit CSU LSU PET RFG -1 -1 -1 -1 Climate change kg CO2 eq 7.42*10 6.00*10 3.26*10 1.02*10 Ozone depletion kg CFC-11 6.33*10-8 5.10*10-8 6.76*10-8 9.65*10-9 eq -3 -3 -3 -4 Terrestrial kg SO2 eq 5.43*10 4.37*10 1.74*10 6.57*10 acidification Freshwater kg P eq 2.54*10-5 2.09*10-5 9.66*10-6 4.80*10-6 eutrophication Marine kg N eq 5.10*10-4 4.13*10-4 2.75*10-4 7.64*10-5 eutrophication Human toxicity kg 1,4-DB 8.80*10-2 7.15*10-2 3.62*10-2 1.20*10-2 eq Photochemical kg NMVOC 4.28*10-3 3.45*10-3 1.20*10-3 5.85*10-4 oxidant formation -3 -3 -4 -4 Particulate matter kg PM10 eq 1.82*10 1.47*10 5.17*10 2.40*10 formation Terrestrial kg 1,4-DB 1.44*10-4 1.16*10-4 2.93*10-5 1.51*10-5 ecotoxicity eq Freshwater kg 1,4-DB 1.09*10-3 9.10*10-4 1.49*10-3 2.83*10-4 ecotoxicity eq Marine ecotoxicity kg 1,4-DB 1.33*10-3 1.10*10-3 1.19*10-3 3.00*10-4 eq Ionising radiation kg U235 eq 5.35*10-2 4.37*10-2 9.42*10-2 9.41*10-3 Agricultural land m²a 1.35*10-1 1.09*10-1 6.30*10-2 1.42*10-2 occupation Urban land m²a 5.67*10-3 4.59*10-3 2.94*10-3 8.31*10-4 occupation Natural land m² 1.69*10-4 1.37*10-4 4.99*10-5 2.81*10-5 transformation Water depletion m³ 1.64*10-3 1.35*10-3 1.66*10-3 5.42*10-4 Metal depletion kg Fe eq 7.56*10-2 7.11*10-2 1.49*10-2 5.54*10-2 Fossil depletion kg oil eq 1.99*10-1 1.61*10-1 1.33*10-1 3.07*10-2

A3.5.2 Municipal scale

Table A3.29 Midpoint level environmental impacts from the municipal scale 2008 reference and alternative packaging scenarios under the ReCiPe (H) LCIA method.

Impact category Unit 2008 reference scenario Alternative packaging scenario 7 7 Climate change kg CO2 eq 2.93*10 1.69*10 Ozone depletion kg CFC-11 eq 2.49 1.49 5 5 Terrestrial acidification kg SO2 eq 2.09*10 1.19*10 Freshwater eutrophication kg P eq 9.79*102 5.91*102 Marine eutrophication kg N eq 1.99*104 1.19*104 Human toxicity kg 1,4-DB eq 3.40*106 1.95*106 Photochemical oxidant kg NMVOC 1.65*105 9.48*104 formation 4 4 Particulate matter formation kg PM10 eq 7.11*10 4.11*10 Terrestrial ecotoxicity kg 1,4-DB eq 5.45*103 2.99*103 Freshwater ecotoxicity kg 1,4-DB eq 4.56*104 3.10*104 Marine ecotoxicity kg 1,4-DB eq 5.43*104 3.51*104 Ionising radiation kg U235 eq 2.13*106 1.36*106 Agricultural land occupation m²a 5.35*106 3.18*106 Urban land occupation m²a 2.24*105 1.36*105 Natural land transformation m² 6.86*103 4.20*103 Water depletion m³ 6.29*104 4.57*104 Metal depletion kg Fe eq 1.13*107 1.01*107 Fossil depletion kg oil eq 7.74*106 4.60*106 245

A3.6 Questionnaires and research consent

The initial purpose of the questionnaires was to acquire information on the packaging and container transportation processes, with specific objectives identified at the beginning of each questionnaire. However, the only responses that have been used from the questionnaires pertain to the geographic origins of the containers. These data were used to estimate the average distances that the containers were transported from the manufacturer to the packager by ship, truck and train. Questionnaires and follow-up communications were sent by mail and e-mail to the wine/spirit companies (to the attention of the operations manager), from 2008 to 2010. Wine/spirit companies receiving the questionnaires were selected based whether or not they supplied wines/spirits to the Ontario market. The questionnaires were limited to the top 100 wine and spirit companies, based on 2008 volume sales, as well as smaller wineries in the Province of Ontario. Questionnaires were sent to 106 wine companies and 28 spirits companies.

246

A3.6.1 Set 1

A3.6.1.1 Wine companies

RESEARCH CONSENT FORM Wineries

Title of Research: Integrating waste prevention activities into life cycle assessments of residential solid waste management: the life cycles of alternative packaging systems for wine and liquor and their potential effects upon the waste management system of Toronto, Canada

Name of Researcher: Julian Cleary, PhD Candidate, Department of Geography, University of Toronto, Sidney Smith Hall, 100 Saint-George St., Rm. 5047, Toronto (Ontario), M5S 3G3, Fax: (416) 946-3886, Tel: (416) 465-1822; e-mail: [email protected]

Research Supervisor: Professor Virginia Maclaren, Department of Geography, University of Toronto, Tel: (416) 978-3375, e-mail: [email protected]

Purpose of the Research: My research focuses on how to integrate waste prevention and product reuse activities into life cycle assessments (LCAs) of solid waste management. Life cycle assessment models of waste management systems have omitted the roles of waste prevention and product reuse in reducing waste generation. I would argue that this limits such LCAs to the evaluation of the environmental effects of waste treatment, not waste management.

Description of the Research: My research explores how a life cycle assessment of municipal solid waste can be capable of evaluating the use of the entire waste hierarchy (waste prevention, product reuse, recycling, biological treatment, thermal treatment and landfilling) in designing waste management systems that would minimize environmental emissions and impacts. In order to evaluate some of the potential implications of an altered life cycle methodology on the design of solid waste management systems, I am undertaking a case study of the effects of wine and liquor packaging alternatives on waste management in the City of Toronto.

Confidentiality: The information obtained through the questionnaires will be combined with data from other wineries that supply wine to the Ontario market and will not be traceable to your establishment.

published, I will destroy the hard copies of the filled-in questionnaires, making it impossible to link the names of the respondents and the organizations that they represent with the information supplied.

Disposition of Research Results: It is my intention to publish the results of this research in academic journals in order to contribute to the development of life cycle assessment methodology for waste management. You will be sent a summary of the results of the study within about one year from receipt of your answers to the questionnaire. A copy of my PhD thesis will be held by the University of Toronto and the National Library of Canada, and will be accessible through either library.

Right of Exclusion or Withdrawal: You have the right to refuse to participate in the research, the right not to answer specific questions, and the right to withdraw from the project at any time up until 1 December 2008.

Consent to Participate: I have read and understood the above consent form. I agree to take part in the above described study.

DATE SIGNATURE

248

QUESTIONNAIRE: WINERIES

Objectives of Questionnaire:

(1) Determine the amount of fuel used during the transportation of the empty wine containers to the wineries. (2) Determine the amount of fuel used during the transportation of the wine to the packaging facilities. (3) Determine the amount of fuel used during the transportation of the filled containers to the Ontario market. (4) Determine the amount of energy and water used to package the wine. (5) Identify the material inputs for the packaging of wine. (6) Determine the amount of solid waste produced from the packaging of wine.

Confidentiality:

Neither the information, nor its source will be identified. The information will be combined with data from other wineries and will not be traceable to your establishment.

Note:

Please note that all of the questions pertain to the portion of your operations that are associated with the production and packaging of WINES (please exclude liquors, coolers and beers). Please leave blank those sections that do not pertain to your company.

249

QUESTIONS

Section A: Production

(1) Please indicate the volume of wine that your company supplied to the Canadian market in 2007.

______Litres ______% Consumed in Ontario

Section B: Filling of Containers

(1) Please indicate the amount of energy (fuel and electricity) used to operate the equipment to rinse, sterilise, fill, cork/cap, label, palletise and shrink wrap the containers. Please specify if the data are per bottle/package, hour, or litre filled.

Operation Electricity Use Unit Fuel Fuel Use Unit (type)

Rinsing ______Sterilisation ______Filling ______Cork/cap ______Labelling ______Palletising ______Shrink wrapping ______OR A combination of the above operations ______

(2) Please indicate the amount of water used during the rinsing and sterilisation of the containers and what becomes of it.

Input of Water

______litres of water / 750ml bottle

Use of the water (e.g., rinsing) and the percentage reused ______

______

250

(3) Please indicate the amount of electricity and/or fuel used to heat the water that is used during the container rinsing and/or sterilisation process.

Electricity ______kWh/litre Natural Gas ______m³/litre

_____ % of heat that is recovered

If recovered, how is this heat used? ______

Section C: Primary, Secondary and Tertiary Packaging

(1) Please indicate the average mass of each wine container that your company supplied to the Canadian market in 2007.

Volume of Container Type of Container Average Mass (g) Number (Glass, PET, aseptic Supplied carton, bag-in-box) ______ml ______g ______ml ______g ______ml ______g ______ml ______g ______ml ______g ______ml ______g ______ml ______g ______ml ______g ______ml ______g ______ml ______g ______

(2) Please indicate the type(s) of pallets (e.g., wooden, HDPE) that are used in the transportation of the wine from the bottler to the retailer.

Type of pallet % of total used Approximate # Mass of empty pallet of times reused ______

(3) Please indicate the average number of bottles/containers that can be loaded onto one pallet. Type and Size of Bottle/Container Glass ______750 ml ____ ml ____ ml ____ ml

Number of Containers ______251

(4) Please indicate the average number of filled bottles/containers that are shipped in one truck, railcar or shipping container.

Truck Railcar Shipping Container Containers (type ______size ___ ml) ______Containers (type ______size ___ ml) ______Containers (type ______size ___ ml) ______Containers (type ______size ___ ml) ______

Section D: Finished Product

(1) Please indicate the percentage of closures used in 2007 that were composed of cork, plastics and aluminium.

% Location of manufacture Unprocessed cork ______Processed cork ______Plastic stopper ______Plastic screw top ______Aluminium screw top ______Other ______

(2) Please indicate the composition and amount of solid waste generated during the packaging process in 2007 and what became of it (e.g., where it was shipped).

Composition of Solid Waste Amount (tonnes) What became of it?

______252

Section E: Transportation Distances Between Wineries and Bottlers

(1) Please identify the location(s) where the wine that your company supplied to the Canadian market was produced in 2007.

Locations (city, country) % of Your Company’s Canadian Sales (by volume)

Producer 1 ______% Producer 2 ______% Producer 3 ______% Producer 4 ______% Producer 5 ______% Producer 6 ______% Producer 7 ______% Producer 8 ______% Producer 9 ______% Producer 10 ______% 100%

(2) Please identify, for each wine production facility [signified by a number in Section B, Question 1]:

(a) The location(s) of the bottlers, and the percentage of wine, by volume, sent to each (b) The source(s) of the bottles/containers and the types of bottles/containers supplied (c) The type of container used to transport the wine in bulk to the bottler (e.g., 22 000 litre single trip polyethylene flexitanks, 24,000 litre single trip polyethylene flexitanks, or 26,000 litre steel ISO tanks) (d) the method(s) of shipment of the wine to the bottler [truck, train, ship, air freight]. (e) the method(s) of shipment of the wine from the bottler to the Ontario (Canada) market [truck, train, ship, air freight].

253

Producer (Wine production facility) 1 a) Location(s) of bottler(s) and the percentage sent to each

Bottler 1 ______% Bottler 2 ______% Bottler 3 ______% b) Source(s) of bottles/packages (city, country) and the types of packages (glass, PET, aseptic cartons, bag-in-box) supplied Source of Packages Types of Packages % of Packages Supplied (by Volume of Wine Packaged) Supplier 1 i ______% ii ______% iii ______% 100% Supplier 2 i ______% ii ______% iii ______% 100% Supplier 3 i ______% ii ______% iii ______% 100% c) Concentration of wine (% in relation to final volume) sent to each bottler and type of container used to transport the wine in bulk to the bottler

Concentration Bulk Shipping Container Type Bottler 1 ___% ______or N/A [ ] Bottler 2 ___% ______or N/A [ ] Bottler 3 ___% ______or N/A [ ] d) Mode of shipment of wine to each bottler and % of the total distance transported

Bottler 1 Truck ___% Train ___% Ship ___% Air ___% Bottler 2 Truck ___% Train ___% Ship ___% Air ___% Bottler 3 Truck ___% Train ___% Ship ___% Air ___% e) Mode of shipment of packaged wine from each bottler to the Ontario, Canada market and % of the total distance transported

Bottler 1 Truck ___% Train ___% Ship ___% Air ___% Bottler 2 Truck ___% Train ___% Ship ___% Air ___% Bottler 3 Truck ___% Train ___% Ship ___% Air ___% 254

Producer (Wine production facility) ___ (please fill in additional copies, if required) a) Location(s) of bottler(s) and the percentage sent to each

Bottler 1 Truck ___% Train ___% Ship ___% Air ___% Bottler 2 Truck ___% Train ___% Ship ___% Air ___% Bottler 3 Truck ___% Train ___% Ship ___% Air ___% 255

Comments and Explanatory Notes

______

Thank you for your time.

256

A3.6.1.2 Spirit companies

RESEARCH CONSENT FORM Distilleries

Title of Research: Waste Prevention and Product Reuse in a Life Cycle Assessment of Residential Waste Management: The WasteMAP Life Cycle Assessment

Name of Researcher: Julian Cleary, PhD Candidate, Department of Geography, University of Toronto, Sidney Smith Hall, 100 Saint-George St., Rm. 5047, Toronto (Ontario), M5S 3G3, Canada, Fax: (416) 946-3886, Tel: (416) 465-1822; e-mail: [email protected]

Research Supervisor: Professor Virginia Maclaren, Department of Geography, University of Toronto, Tel: (416) 978-3375, e-mail: [email protected]

Confidentiality: The information obtained through the questionnaires will be combined with data from other distilleries that supply spirits to the Ontario market and will not be traceable to your establishment.

I will store the answers to the questionnaires in a locked filing cabinet and the results on my desktop computer (the names of the respondents and the organizations that they represent will not be linked with the data in the spreadsheet file). Once the research is published, I will destroy the hard copies of the filled-in questionnaires, making it impossible to link the names of the respondents and the organizations that they represent with the information supplied. 257

Consent to Participate: I have read and understood the above consent form. I agree to take part in the above described study.

DATE SIGNATURE

258

QUESTIONNAIRE: DISTILLERIES

Objectives of Questionnaire:

(1) Determine the amount of fuel used during the transportation of the empty liquor containers to the distilleries. (2) Determine the amount of fuel used during the transportation of the liquor to the bottling facilities. (3) Determine the amount of fuel used during the transportation of the filled containers to the Ontario market. (4) Determine the amount of energy and water used to package the liquor. (5) Identify the material inputs for the packaging of the liquor. (6) Determine the amount of solid waste produced from the packaging of liquor.

Confidentiality:

Neither the information, nor its source will be identified. The information will be combined with data from other distilleries and will not be traceable to your establishment.

Note:

Please note that all of the questions pertain to the portion of your operations that are associated with the production and packaging of LIQUOR/SPIRITS (please exclude wines, coolers and beers). Please leave blank those sections that do not pertain to your company.

259

QUESTIONS

Section A: Production

(1) Please indicate the volume of liquor that your company supplied to the Canadian market in 2007.

______Litres _____ % Consumed in Ontario

Section B: Filling of Containers

(1) Please indicate the amount of energy (fuel and electricity) used to operate the equipment to rinse, sterilise, fill, cork/cap, label, palletise and shrink wrap the containers (excluding the energy to heat the spirit, if necessary). Please specify if the data are per bottle/package, hour, or litre filled.

Operation Electricity Use Unit Fuel Fuel Use Unit (type)

Rinsing ______Sterilisation ______Filling ______Cork/cap ______Labelling ______Palletising ______Shrink wrapping ______OR A combination of all the above operations ______

(2) Please indicate the amount of water used during the rinsing and sterilisation of the containers and what becomes of it.

Input of Water

______litres of water / litre filled

Use of the water (e.g., rinsing) and the percentage reused ______

______

260

(3) Please indicate the amount of electricity and/or fuel used to heat the water that is used during the container rinsing and/or sterilisation process.

Electricity ______kWh/litre Natural Gas ______m³/litre

_____ % of heat that is recovered

If recovered, how is this heat used? ______

Section C: Primary, Secondary and Tertiary Packaging

(1) Please indicate the average mass of each liquor container that your company supplied to the Canadian market in 2007.

Volume of Container Type of Container Average Mass (g) Number (Glass, PET, other) Supplied ______ml ______g ______ml ______g ______ml ______g ______ml ______g ______ml ______g ______ml ______g ______ml ______g ______ml ______g ______ml ______g ______ml ______g ______

(2) Please indicate the type(s) of pallets (e.g., wooden, HDPE) that are used in the transportation of the liquor from the bottler to the retailer.

Type of pallet Approximate # of times reused Mass of empty pallet ______

(3) Please indicate the average number of bottles/containers that can be loaded onto one pallet. Type and Size of Bottle/Container

Glass ______750 ml ____ ml ____ ml ____ ml

Number of Containers ______261

(4) Please indicate the average number of filled bottles/containers that are shipped in one truck, railcar or shipping container.

Truck Railcar Shipping Container Containers (type ______size ___ ml) ______Containers (type ______size ___ ml) ______Containers (type ______size ___ ml) ______

Section D: Finished Product

(1) Please indicate the percentage of closures used in 2007 that were composed of cork, plastics and aluminium. % Location(s) of manufacture Unprocessed cork (plastic top) ______Processed cork (plastic top) ______Plastic stopper ______Plastic screw top ______Aluminium screw top ______Other ______

(2) Please indicate the composition and amount of solid waste generated during the packaging of liquor in 2007 and what became of it (e.g., where it was shipped).

Composition of Solid Waste Amount (tonnes) What became of it?

______

262

Section E: Transportation Distances Between Distilleries and Bottlers

(1) Please identify the location(s) where the liquor that you supplied to the Canadian market was produced in 2007.

Locations (city, country) % of Your Company’s Canadian Sales (by volume) Producer 1 ______% Producer 2 ______% Producer 3 ______% Producer 4 ______% Producer 5 ______% Producer 6 ______% Producer 7 ______% Producer 8 ______% Producer 9 ______% Producer 10 ______% 100%

(2) Please identify, for each liquor production facility [signified by a number in Section B, Question 1]:

(a) The location(s) of the bottlers, and the percentage of liquor, by volume, sent to each (b) The source(s) of the bottles and the types of bottles supplied (c) Whether or not the liquor was sent to the bottler in concentrated form, the average level of concentration, the type of container used to transport the liquor in bulk to the bottler (e.g., 22 000 litre single trip polyethylene flexitanks, 24,000 litre single trip polyethylene flexitanks, or 26,000 litre steel ISO tanks) (d) the method(s) of shipment of the liquor to the bottler [truck, train, ship, air freight]. (e) the method(s) of shipment of the liquor from the bottler to the Ontario (Canada) market [truck, train, ship, air freight].

263

Producer (Liquor production facility) 1

a) Location(s) of bottler(s) and the % of liquor shipments, by volume, sent to each

Bottler 1 ______% Bottler 2 ______% Bottler 3 ______%

b) Source(s) of bottles (city, country) and the types of bottles (glass, PET) supplied Source of Bottles Types of Bottles % of Bottles Supplied (by Volume of Liquor Packaged) Supplier 1 i ______% ii ______% iii ______% 100% Supplier 2 i ______% ii ______% iii ______% 100% Supplier 3 i ______% ii ______% iii ______% 100%

c) Concentration of liquor (% in relation to final volume) sent to each bottler and type of container used to transport the liquor in bulk to the bottler

Concentration Bulk Shipping Container Type Bottler 1 ___% ______or N/A [ ] Bottler 2 ___% ______or N/A [ ] Bottler 3 ___% ______or N/A [ ]

d) Mode of shipment of bulk liquor to each bottler and % of the total distance transported

Bottler 1 Truck ___% Train ___% Ship ___% Air ___% Bottler 2 Truck ___% Train ___% Ship ___% Air ___% Bottler 3 Truck ___% Train ___% Ship ___% Air ___%

e) Mode of shipment of packaged liquor from each bottler to the Ontario, Canada market and % of the total distance transported

Bottler 1 Truck ___% Train ___% Ship ___% Air ___% Bottler 2 Truck ___% Train ___% Ship ___% Air ___% Bottler 3 Truck ___% Train ___% Ship ___% Air ___% 264

Producer (Liquor production facility) ___ (please fill in additional copies, if required) a) Location(s) of bottler(s) and the % of liquor shipments, by volume, sent to each

Bottler 1 ______% Bottler 2 ______% Bottler 3 ______% b) Source(s) of bottles (city, country) and the types of bottles (glass, PET) supplied Source of Bottles Types of Bottles % of Bottles Supplied (by Volume of Liquor Packaged) Supplier 1 i ______% ii ______% iii ______% 100% Supplier 2 i ______% ii ______% iii ______% 100% Supplier 3 i ______% ii ______% iii ______% 100% c) Concentration of liquor (% in relation to final volume) sent to each bottler and type of container used to transport the liquor in bulk to the bottler

Concentration Bulk Shipping Container Type Bottler 1 ___% ______or N/A [ ] Bottler 2 ___% ______or N/A [ ] Bottler 3 ___% ______or N/A [ ]

d) Mode of shipment of bulk liquor to each bottler and % of the total distance transported

Bottler 1 Truck ___% Train ___% Ship ___% Air ___% Bottler 2 Truck ___% Train ___% Ship ___% Air ___% Bottler 3 Truck ___% Train ___% Ship ___% Air ___%

e) Mode of shipment of packaged liquor from each bottler to the Ontario, Canada market and % of the total distance transported

Bottler 1 Truck ___% Train ___% Ship ___% Air ___% Bottler 2 Truck ___% Train ___% Ship ___% Air ___% Bottler 3 Truck ___% Train ___% Ship ___% Air ___% 265

Comments and Explanatory Notes

______

Thank you for your time.

266

A3.6.2 Set 2

A3.6.2.1 Wine companies

RESEARCH CONSENT

Title of Research: Incorporating waste prevention activities into life cycle assessments of residential solid waste management

Research Supervisor: Professor Virginia Maclaren, Department of Geography, University of Toronto, Tel: (416) 978-3375, e-mail: [email protected]

Purpose of the Research: My research focuses on how to integrate waste prevention activities, such as alternative packaging for wine and liquor, into life cycle assessments (LCAs) of residential solid waste management. I am undertaking a case study of the effects of wine and liquor packaging systems on residential solid waste management in the City of Toronto.

Confidentiality: The information provided by the research participant will be combined with data from the other participants. In any oral or written presentation of the results of this study, the names of participants will NOT be identified, nor will the source of the data.

Once receiving the e-mail responses, I will copy the requested data to a Microsoft Excel file on my desktop computer. The names of the respondents and the organizations that they represent will not be linked with the data in the file. Immediately after copying the data from the e-mails, I will delete all of the e-mail responses from the computer.

Disposition of Research Results: It is my intention to publish the results of this research in academic journals in order to contribute to the development of life cycle assessment methodology for waste management. A copy of my PhD thesis will be held by the University of Toronto and the National Library of Canada in Ottawa and will be accessible through these libraries.

Right of Exclusion or Withdrawal: You have the right to refuse to participate in the research and the right to withdraw from the project at any time up until 1 March 2010.

Consent to Participate: I have read and understood the above consent form. I agree to take part in the above described study. 267

QUESTIONS

1) Please specify the approximate volume of wine your company supplied to the Ontario, Canada market in 2008 (if only 2009 data is available, please specify).

______litres

2) Please indicate the geographic origin (country and city) of the various types of containers used to package your wine (destined for the Ontario market, if possible).

ANSWER FORMAT: % of containers of each type, origin of containers, location filled

EXAMPLE OF ANSWER: -75% of our 750 ml glass containers are manufactured in (city, country) and shipped to (city, country) for filling -25% of our 750 ml glass containers are manufactured in (city, country) and shipped to (city, country) for filling -100% of our 1 litre aseptic cartons (tetra paks) are manufactured in (city, country) and shipped to (city, country) for filling

3) Please specify the % of the glass bottles used to package your wine that is clear glass and coloured glass.

ANSWER FORMAT: % clear glass, % coloured glass

268

A3.6.2.2 Spirit Companies

RESEARCH CONSENT

Title of Research: Incorporating waste prevention activities into life cycle assessments of residential solid waste management

Research Supervisor: Professor Virginia Maclaren, Department of Geography, University of Toronto, Tel: (416) 978-3375, e-mail: [email protected]

Disposition of Research Results: It is my intention to publish the results of this research in academic journals in order to contribute to the development of life cycle assessment methodology for waste management. A copy of my PhD thesis will be held by the University of Toronto and the National Library of Canada in Ottawa and will be accessible through these libraries.

Right of Exclusion or Withdrawal: You have the right to refuse to participate in the research and the right to withdraw from the project at any time up until 1 March 2010.

Consent to Participate: I have read and understood the above consent form. I agree to take part in the above described study.

269

QUESTIONS

1) Please indicate the geographic origin (country and city) of the various types of containers used to package your liquor/spirits (destined for the Ontario, Canada market in 2008, if possible).

ANSWER FORMAT: % of containers of each type, origin of containers, location filled

2) Please specify the % of the glass bottles used to package your liquor/spirits that is clear glass and coloured glass.

ANSWER FORMAT: % clear glass, % coloured glass

3) If known, please indicate the average mass of the containers that your company uses for its liquor/spirits.

ANSWER FORMAT: Volume and type of container, mass of container

EXAMPLE OF ANSWER: 750 ml glass container: 550 g 1140 ml glass container: 850 g

270

3.7 References

Encyclopedia of Earth. 2010. Aluminum. Available at: http://www.eoearth.org/article/Aluminum. Accessed on 2011 05 27.

Google. 2010. Google Maps. Available at: http://maps.google.ca/maps. Accessed on 2010 02 18.

Integrated Design Engineering Systems (IDES). n.d. Polyethylene (PE) Plastic. Available at: http://www.ides.com/info/generics/27/Polyethylene-PE. Accessed on 2011 05 27.

Koel, J. 2008. Personal communication with J. Koel. Tetra Pak Canada. Richmond Hill, Ontario, Canada, 2008 12 15.

Stewart, T. 2010. Personal communication with T. Stewart. Liquor Control Board of Ontario. Toronto, Ontario, Canada. 2010 07 27.

Statistics Canada. 2008. CANSIM. Table 051-0046. Catalogue no. 91C0029.

World-Ports Distances Calculator. 2010. Available at http://www.distances.com/. Accessed on 2010 08 07.

271

Appendix 4

Paper 4

272

A4.1 Field research

A4.1.1 Admail WPA

The field research for the admail WPA was undertaken at a single detached house in Toronto during the week of 7-13 February 2011. All of the unaddressed advertising mail delivered by private carriers (excluding the admail within subscription newspapers) was weighed using a kitchen scale. In order to extrapolate the measured mass of admail received at one household to the amount received at all of Toronto’s 996,893 households in Toronto in 2008 (WDO 2008), the following procedure was employed: (1) the measured mass of admail in one week was multiplied by 52 (the number of weeks in one year); (2) the result was then multiplied by the number of households in Toronto. From these calculations, it was estimated that Toronto households receive approximately 1.53*104 tonnes of unaddressed advertising mail per year (excluding the amount delivered by postal carriers as well as admail in subscription newspaper inserts). The admail measurements are listed in Table A4.1.

Table A4.1 Mass of unaddressed admail received at a single detached Toronto household during the week of 7-13 February 2011

Type of Mass of unaddressed advertising mail delivered (g) paper 7 Feb 2011 8 Feb 2011 9 Feb 2011 10 Feb 2011 11 Feb 2011 12 Feb 2011 13 Feb 2011 Glossy 20 17 0 37 0 0 0 Newsprint 16 244 0 260 0 0 0 Total 36 261 0 297 0 0 0

A4.1.2 Newspaper WPA

An estimate of the mass of daily subscription newspapers read in the City of Toronto in 2008 was produced by weighing a sample of four newspapers (Toronto Star, The Globe and Mail, National Post, and Toronto Sun) each day from Monday, January 17th to January 23rd 2011, using a digital kitchen scale. The mass measurements are listed in Table A4.2.

273

Table A4.2 Measured mass of daily subscription newspapers in Toronto during the week of 17-23 January 2011

Newspaper Mass of daily subscription newspapers (g) 17 January 18 January 19 January 20 January 21 January 22 January 23 January 2011 2011 2011 2011 2011 2011 2011 Toronto Star 172 206 220 346 255 743 197 The Globe 199 222 245 228 357 504 0 and Mail Toronto Sun 159 164 179 226 230 177 254 National 166 168 183 195 239 365 0 Post

The mass of each newspaper (Table A4.2) was multiplied by the total number of newspapers of the particular publication that were sold on the given day of the week in 2008 (Table A4.3). Based on the circulation figures from Canadian Newspaer Association (2009), 3.480*108 daily subscription newspapers were sold in the Greater Toronto Area market in 2008. The amount sold in the City of Toronto is estimated at 46.17% of this figure, based on the percentage of the population of the census metropolitan area of Toronto that resides in the City of Toronto (City of Toronto 2006, Statistics Canada 2008a, 2006). From the 2008 sales in the City of Toronto, it is assumed that 10% (1.607*107 newspapers, equal to 4,528 tonnes of newsprint) is prevented under the waste prevention scenario.

Table A4.3 2008 daily circulation of subscription newspapers in Toronto based on statistics from the Canadian Newspaper Association (2009)

Newspaper Circulation of daily subscription newspapers in Toronto in 2008 Monday Tuesday Wednesday Thursday Friday Saturday Sunday Toronto Star 309,793 307,019 306,395 307,197 310,169 477,683 331,504 The Globe 321,109 321,109 321,109 321,109 321,109 391,040 0 and Mail Toronto Sun 154,602 156,570 157,399 159,430 154,513 133,452 246,898 National 194,795 194,795 194,795 194,795 194,795 208,234 0 Post

In the waste prevention scenario, the service that would have been provided by the prevented newspapers was supplied through the use of computers to read newspaper content online. Based on the data and assumptions described in Section 5.5.1.3 of Paper 4, it was estimated that: (1) there were 51.1 million readers for the prevented newspapers, based on the estimate of 3.18 readers per newspaper copy (Scarborough Research and Newspaper National Network LP 2010); (2) the time spent reading these 274

newspapers was substituted with 17 million hours spent reading newspaper articles online (20 minutes on the computer, per reader); (3) 260 million MB of data were downloaded onto computers (based upon 5 MB of downloaded content per reader each day); (4) the electricity used to download the data was 1.8 million kWh. Note: numbers may not add up due to rounding.

A4.2 Residential waste calculations

A4.2.1 Composition of residual waste

Although data on the composition of the marketed recyclables from residential waste is available from Waste Diversion Ontario (WDO), data on the composition of the residual waste (i.e., the waste to be landfilled) is not. The results from waste audits of single family and multiple unit households in Toronto’s downtown and Scarborough neighbourhoods in 2005 and 2006 were extrapolated to generate an estimate of the composition of the residual (landfilled) waste in 2008. The residential waste audits, undertaken by the City of Toronto through the Stewardship Ontario Waste Audit Program, had a sample size of 100 households which were monitored over one season, comprising 13 weeks (Stewardship Ontario 2006a, 2006b). The proportions of each waste stream were derived from this audit data and assumed representative of the residual waste from all of the single family dwellings (SFDs) and multiple family dwellings (MFDs) in the entire city. The Solid Waste Management Services Division of the City of Toronto estimated that 50.05% of the dwellings were SFDs while 49.95% were MFDs in the year 2006 (Waste Diversion Ontario 2008).

A4.2.2 Composition of recyclable waste

Estimates of the composition and amount of recyclables collected from residential sources were calculated from statistics supplied by the City of Toronto and the WDO (2008). The former publishes the amount of residential recyclables marketed annually (not broken down by composition), while the latter supplies annual data on the 275 composition and amount of recyclables collected not only from residential sources, but also schools and public spaces. The amount of recyclables stated in the WDO figure is 5.0% greater than the City of Toronto figure because the former does not isolate residential recyclables from the other two sources of recyclables. Therefore, data was normalised by multiplying the percentage composition of each type of recyclable in the WDO data, by the total mass of residential recyclables marketed by the City of Toronto. Due to the miniscule contribution of schools and public spaces to the amount of recyclables collected by the municipality, it is assumed that these sources do not generate a significant effect on the overall composition of the recycled material collected. Since the waste audits address those circumstances from 2005 and 2006, the results would exclude the effects of the introduction of the wine and spirit container deposit-return system in February 2007 on recycling rates. Although these waste audits have “LCBO (wine and spirit container) glass” and “Other glass” categories for landfilled waste, the WDO municipal glass recycling figure addresses all glass in one category. In order to generate an estimate of the effects of the deposit-return system on the municipal recycling rates of the “LCBO glass” and “Other glass” categories, the following assumptions were made: (1) “LCBO glass” is the amount of wine and spirit glass containers not collected through the provincially-mandated deposit-return system minus amount of LCBO glass landfilled; (2) “Other glass” is the amount of all glass municipally recycled minus amount of LCBO glass municipally recycled. The waste audits do not isolate aseptic carton and plastic bottle packaging for wines and spirits from their respective waste categories (polycoat and PET). Therefore, it is assumed that the deposit-return system has no effect on the municipal recycling rate of polycoat and PET waste.

A4.3 Transportation data, assumptions and calculations

The transport distances by land to and from the City of Toronto are estimated using Google Maps (Google 2010) (first option selected).

276

A4.3.1 Transportation of newsprint

The 541 km estimate of the average transportation distance of newsprint from the paper mill to Toronto was based upon the average distance from Toronto to the three sources of newsprint for the Toronto Star, the daily newspaper with the largest circulation in Canada. According to this newspaper, the newsprint is derived from mills in Trois- Rivières, Quebec (673 km), Kapuskasing, Ontario (831 km) and Thorold, Ontario (120 km) (Toronto Star n.d.). From an analysis of Statistics Canada 2003 and 2004 data undertaken by Travacon Research Limited (2007) in its report for the Forest Products Association of Canada, it was estimated that 75% of Canadian pulp and paper is shipped by rail, with the remainder by truck. This modal split is used for calculating the emissions from the transportation of newsprint.

A4.3.2 Residential waste transport

Table A4.4 lists the residential waste treatment/disposal/transfer locations and distances from Toronto, mass collected and shipped, and tonne-km calculated for shipments to treatment/disposal/transfer locations in 2008.

Table A4.4 Residential waste treatment/disposal locations and estimated distances from Toronto, and tonne-km calculated for shipments to treatment/disposal locations in 2008

Type of Treatment / Distance Tonne-km waste disposal / transfer from 2008 reference scenario Waste prevention scenario location Toronto (km) Waste collection All MSW One of seven MSW 19.8 km 1.64*107 1.60*107 collected transfer facilities one-way trip1 Biological treatment SSO waste Toronto, Ontario 0 0 0 Saint-Henri-de- 806 2.75*107 2.75*107 Levis, Quebec London, Ontario 191 3.82*106 3.82*106 Newmarket, Ontario 54 5.09*105 5.09*105 Tracy, Quebec 621 1.46*106 1.46*106 Courtice, Ontario 70 1.65*105 1.65*105 Welland, Ontario 133 3.13*105 3.13*105 Yard waste Arthur, Ontario 116 3.46*106 3.29*106 / Christmas Thorold, Ontario 133 2.97*106 2.83*106 trees Markham, Ontario 37 8.27*105 7.87*105 London, Ontario 191 1.11*106 1.05*106 Hornby, Ontario 48 7.95*104 7.56*104 Newmarket, Ontario 54 4.47*104 4.25*104 277

Landfill Residual Carleton Farms 441 1.94*108 1.90*108 waste Landfill, Michigan Green Lane 203 6.50*106 6.39*106 Landfill, Ontario2 Municipal recycling Newsprint Whitby, Ontario 49 1.89*106 1.69*106 Thorold, Ontario 133 5.13*106 4.58*106 China (Shanghai) 13700 4.88*107 4.35*107 (rail) (rail) 1.09*108 9.74*107 (ship) (ship) Cardboard Toronto, Ontario 20 2.29*105 2.29*105 China (Shanghai) 13700 2.18*107 2.18*107 (rail) (rail) 4.87*107 4.87*107 (ship) (ship) Buffalo, New York 165 5.22*105 5.22*105 Mixed Ontario (regional 150 6.33*105 6.33*105 paper distance) China (Shanghai) 13700 3.80*107 3.80*107 (rail) (rail) 8.49*107 8.49*107 (ship) (ship) Glass Belleville, Ontario 130 1.01*106 9.48*105 Brampton, Ontario 43 2.22*105 2.07*105 Stanley, New York 318 1.19*106 1.11*106 Montreal, Quebec 545 1.74*106 1.63*106 Sarnia, Ontario 288 1.02*105 9.58*104 Comingled Oshawa, Ontario 61 3.97*105 3.97*105 Brampton, Ontario 43 3.11*104 3.11*104 Steel Hamilton, Ontario 70 3.33*105 3.33*105 PET Montreal, Quebec 545 1.27*106 1.28*106 Winnipeg, Manitoba 2050 1.43*106 1.43*106 HDPE Sarnia, Ontario 288 5.79*105 5.79*105 Montreal,Quebec 545 4.56*104 4.56*104 Aluminum Oswego, New York 414 3.61*105 3.61*105 Hamilton, Ontario 70 1.53*104 1.53*104 Poly-coat Bannockburn, 887 1.47*105 1.47*105 Illonois Inchon, South 13700 2.75*106 2.76*106 Korea (rail) (rail) 5.79*106 5.81*106 (ship) (ship) Plastic tubs Prescott, Ontario 357 1.00*105 1.00*105 and lids Polystyrene Malton, Ontario 32 1.31*103 1.31*103 LDPE Elmira, Ontario 126 1.20*103 1.15*103 Net deposit-return (The Beer Store) recycling Glass Belleville, Ontario 130 7.61*105 3.95*105 Brampton, Ontario 43 1.67*105 8.65*104 Stanley, New York 318 8.92*105 4.63*105 Montreal, Quebec 545 1.31*106 6.80*105 Sarnia, Ontario 288 7.69*104 3.99*104 Paper Ontario (regional 150 2.28*103 1.53*103 distance) China (Shanghai) 13706 1.37*105 9.15*104 (rail) (rail) 3.06*105 2.05*105 (ship) (ship) PET Montreal, Quebec 545 1.28*104 1.97*104 Winnipeg, Manitoba 2050 1.44*104 2.21*104 Poly-coat Bannockburn, 887 8.95*102 1.77*103 Illonois 278

Inchon, South 13700 1.68*104 3.32*104 Korea (rail) (rail) 3.54*104 6.99*104 (ship) (ship) 1 Average distance supplied by Solid Waste Managemetn Services of the City of Toronto (2010a) 2 Approximate location of Green Lane Landfill in St. Thomas, Ontario: 42.813049 degrees latitude, -81.327202 degrees longitude.

A4.4 Unit processes

A4.4.1 Selected unit processes

Tables A4.5 and A4.6 list the upstream, downstream unit processes selected for the WasteMAP scenarios, the emissions database source, and the year and geography of the technological representation of the processes.

Table A4.5 Unit processes for upstream component of the WasteMAP LCA scenarios

Materials/Processes Unit process description LCA database Year Geography of technology WPA 1 Newsprint Paper, newsprint, at plant US-EI Unspecified Default: Avg. European situation time period -Altered elect. mix WPA 2 HDPE Packaging flim, HDPE, at US-EI Unspecified Default: Avg. European situation plant1 time period -Altered elect. mix WPA 3 Newsprint Paper, newsprint, at plant US-EI Unspecified Default: Avg. European situation time period -Altered elect. mix Computer use Use, computer, desktop US-EI 2007 Default: Avg. European situation with LCD monitor, active -Altered elect. mix mode WPAs 4 and 5 Containers Glass bottles Packaging glass, brown, at US-EI Unspecified Default: Avg. European situation plant/55% Cullet time period -Altered elect. mix Packaging glass, green, at US-EI Unspecified Default: Avg. European situation plant/75% Cullet time period -Altered elect. mix Packaging glass, white, at US-EI Unspecified Default: Avg. European situation plant/58% Cullet time period -Altered elect. mix Aseptic cartons Aluminium, primary, at US-EI Unspecified Avg. European situation plant time period Solid bleached board, SBB, US-EI Unspecified Avg. European situation at plant time period Packaging film, LDPE, at US-EI Unspecified Avg. European situation plant time period Aseptic carton manufacture Author defined 2002 Dijon, France with avg. European unit process electricity PET bottles Polyethylene terephthalate, US-EI Unspecified Default: Avg. European situation granulate, bottle grade, at time period with US electricity plant -Altered elect. mix Blow molding US-EI Unspecified Default: Avg. European situation time period with US electricity -Altered elect. mix Injection molding US-EI Unspecified Default: Avg. European situation time period with US electricity -Altered elect. mix 279

Secondary packaging (capsules, closures and paper labels) Capsules Aluminium, primary, at US-EI Unspecified Avg. European situation with US plant time period electricity Polyethylene terephthalate, US-EI Unspecified Avg. European situation with US granulate, amorphous, at time period electricity plant Polyvinylidenchloride US-EI Unspecified Avg. European situation with US (PVDC), granulate, at plant time period electricity Tin, at regional storage US-EI Unspecified Avg. European situation with US time period electricity Closures Aluminium, primary, at US-EI Unspecified Avg. European situation with US plant time period electricity Kraft paper, bleached, at US-EI 2000 Avg. European situation with US plant electricity Polyethylene, HDPE, US-EI Unspecified Avg. European situation with US granulate, at plant time period electricity Polyethylene, LDPE, US-EI Unspecified Avg. European situation with US granulate, at plant time period electricity Polyethylene terephthalate, US-EI Unspecified Avg. European situation with US granulate, amorphous, at time period electricity plant Polypropylene, granulate, at US-EI Unspecified Avg. European situation with US plant time period electricity Polyvinylidenchloride US-EI Unspecified Avg. European situation with US (PVDC), granulate, at plant time period electricity Raw cork, at forest road EcoInvent 1993 Avg. European situation Tin, at regional storage US-EI Unspecified Avg. European situation with US time period electricity Paper labels for glass Paper, wood-containing, US-EI Unspecified Avg. European situation with US and PET wine and LWC, at regional storage time period electricity spirit containers Transportation Transportation by Transport, lorry >28t, fleet US-EI Unspecified Switzerland with US electricity truck average time period Transportation by rail Transport, freight, rail, US-EI Unspecified United States diesel time period Transportation by Transport, transoceanic US-EI Unspecified Global (extrapolated from a port in ship freight ship time period Netherlands) 1 This unit process was put together by adding the process “Extrusion, plastic flim” to “Polyethylene, HDPE, granulate, at plant.”

Table A4.6 Processes for downstream component of the WasteMAP LCA scenarios (excluding recycling)

Processes Process description LCA database Year Geography of technology Residential waste Transport, municipal waste EcoInvent Unit Unspecified Switzerland with US collection collection, lorry 21t Process time period electricity Recyclable material Recyclables, sorted at MRF, Author defined unit Unspecified Avg. European situation sorting at MRF for further treatment process based upon time period with Ontario electricity (excludes transport) US-EI (original supply mix from 2008 process: waste paper, sorted, for further treatment/RER with US electricity U) Residential waste Transport, freight, rail, US-EI Unspecified United States with US transportation diesel time period electricity Transport, lorry >28t, fleet US-EI Unspecified Switzerland with US average time period electricity Transport, transoceanic US-EI Unspecified OCE (Oceanic) freight ship time period Landfilling Landfill US-EI Unspecified Switzerland with US time period electricity (for US landfill) -Ontario electricity mix for 280

Ontario landfill Source separated SSO anaerobic digestion Author defined unit 2005 Toronto, Canada, with organic (SSO) waste process Ontario electricity supply treatment mix from 2008

Compost, at plant US-EI 1997 Default: Switzerland with US electricity -Altered elect. mix Yard waste treatment Compost, at plant US-EI 1997 Default: Switzerland with US electricity -Altered elect. mix Bottle washing Bottle washing Author defined unit 2010 Global (extrapolated from process data supplied by a wine bottle washing equipment manufacturer)

A4.4.2 Author-defined unit processes

Note: The following author-defined unit processes are described in Appendix 3 (Section A3.3.2: (1) aseptic carton manufacture; (2) bottle washing; (3) glass production and recycling; and (4) PET recycling.

A4.4.2.1 Anaerobic digestion of source-separated organic waste

An author-defined process for the anaerobic digestion of residential organic waste was created using data from the operating records from the City of Toronto’s Dufferin Organics Processing Facility in 2008. Those processes undertaken at the DOPF, which treated approximately 40% of the City’s SSO waste in 2008, were assumed to be representative of those taking place at the other facilities that process Toronto’s SSO waste. The material and energy inputs to the defined process are normalized to the quantity of SSO waste inputs to the treatment facility, including the contaminants.

Source of data: City of Toronto 2010

Specifications of defined process: Known outputs to technosphere: SSO waste treatment; 1 tonne Known inputs from nature (resources): Water, process, unspecified origin/m³; 0.61 m³ Known inputs from technosphere (electricity/heat): Electricity, medium voltage, production UCTE, at grid/Ontario 2008; 50.3 kWh 281

Heat, natural gas, at boiler atm. low-NOx condensing non-modulating <100kw/RER with US electricity U; 267.4 MJ (approximately equivalent to 7.4 m³ of natural gas)

Known outputs to technosphere. Waste and emissions to treatment: Treatment, sewage, to wastewater treatment, class 3/CH with US electricity U; 0.94 m³ Flared biogas; 120 m³ (normalized equivalent to the air emissions from the US-EI unit process entitled: “Heat, at cogen with biogas engine, allocation exergy/CH with US electricity U”)

A4.4.2.2 HDPE recycling

Due to the cut-off allocation adopted in the US-EI database, and the lack of a recent unit process for HDPE recycling, data from a published LCA of plastics recycling were used to formulate an LCA process for the recycling of HDPE. The reprocessing efficiency (92.7%) of HDPE recycling in this LCA was adopted. Environmental burdens associated with surfactants, defoamers, wetting agents and alkaline cleaners needed to be omitted because their chemical compositions were not defined in the LCA, and representative unit processes not available in the LCA databases. Environmental burdens caused by the production of the recycling equipment and facility are excluded from the defined process.

Source of data: Franklin Associates 2010

Specifications of defined process: Known outputs to technosphere: Recycling of HDPE; 1000 lb Known inputs from nature (resources): Water, process and cooling, unspecified natural origin; 53.3 gal Known inputs from technosphere (materials/fuels): Sodium hydroxide, production mix, at plant/kg NREL/RNA; 0.27 lb Electricity, medium voltage, at grid/Ontario 2008; 222 kwh Natural gas in industrial equipment (Franklin Associates 1998); 123 cuft LPG into industrial boilers; 0.094 gal Propane/butane, at refinery/RER with US electricity U; 0.15618 kg Emissions to air: Particulates; 0.038 lb VOC, volatile organic compounds; 0.037 lb Known outputs to technosphere. Waste and emissions to treatment: 282

Treatment, sewage, unpolluted, to wastewater treatment, class 3/CH with US electricity U; 53.3 gal

Emissions to water (excluded from process, listed for information purposes): BOD5, Biological Oxygen Demand; 0.3 lb COD, Chemical Oxygen Demand; 0.015 lb Suspended solids, unspecified; 0.29 lb DOC, Dissolved Organic Carbon; 0.0091 lb

A4.4.2.3 Recyclables, sorted at MRF, for further treatment/Ontario 2008

Sources of data: Franklin Associates 2010 for process energy inputs US-EI unit process entitled “waste paper sorted, for further treatment/RER with US electricity U” for material (excluding wastes) and capital inputs Note: Since the processes associated with the transport of the recyclable materials are already included in each LCA scenario, they are omitted from this process.

Specifications of defined process: Known outputs to technosphere: Recyclables, sorted at MRF, for further treatment/Ontario 2008 Known inputs from technosphere (materials/fuels): Electricity, medium voltage, production UCTE, at grid/Ontario 2008; 16.36 kWh Natural gas, combusted in industrial equipment NREL/US; 0.0022 m³ Diesel, combusted in industrial equipment/Ontario 2008; 1.8 litres Propane (substituted with liquefied petroleum gas, combusted in industrial boiler NREL/US); 2.5 litres Steel converter, unalloyed, at plant/RER with US electricity U; 0.0012 kg Sheet rolling, steel/RER with US electricity U; 0.0012 kg Lubricating oil, at plant/RER with US electricity U; 1.2*10-5 kg Waste paper sorting plant/RER/I with US electricity U; 6.06*10-10 p (pieces)

A4.4.3 Electricity production mixes

A description of the procedure used to substitute more appropriate electricity mixes for the unit processes incorporated into each LCA scenario is located in Appendix 3 (Section A3.3.3).

283

A4.4.4 Recycling, biological treatment, and avoided burdens

Both the EcoInvent and US-EI databases apply the cut-off allocation for recycling, which excludes all recycling processes from the LCA system boundary, including any avoided burdens (Frischknecht et al. 2007). These databases often lack the unit processes for recycling, making it necessary for the user to either define the processes from industry data, or use alternative databases. For certain recycling streams, the documentation for the database provides suggestions for the avoided product and the “input from technosphere.” For others in which there are no suggested processes (or the suggested substitute processes are deemed highly questionable in the database documentation), it was necessary to use older data from the Franklin USA 98 database or define processes based upon a Franklin Associates LCA from 2010. Although the given unit processes in the LCA databases mostly depict recycling in terms of the quantities of new materials generated that include recycled content (e.g., kg of polystyrene produced with 50% recycled content), this LCA represents recycling processes relative to the quantities of recycled material inputs used (e.g., kg of waste polystyrene inputs into the polystyrene production process). It is also necessary to ensure that the recycling burdens and avoided burdens resulting from the displacement of virgin materials are equivalent. For example, the Franklin USA 98 unit process for polystyrene (PS) recycling assumes that 500 lb (249 kg) of waste PS inputs (plus 50 lbs of processing losses from these waste inputs) are required to produce 1000 lb (453 kg) of new PS. Therefore, to depict the recycling of the 41 tonnes of waste the PS in the 2008 reference scenario, the emissions from the production of 75 tonnes of PS material with 50% recycled content are included in the LCA, while the emissions associated with the production of the same amount of virgin PS material are subtracted. Table A4.7 lists the recyclable materials, identifies the unit processes used to depict recycling and the avoided products due to recycling, the source of the unit processes, year of data acquisition and the geography of the technology.

284

Table A4.7 Unit processes for the recycling component of the WasteMAP LCA scenarios

Materials Unit process LCA database Year Geography of % Reprocessing efficiency (% of description Technology waste inputs that are incorporated into the final product) Aluminum Avoided product US-EI Unspecified Avg. European - Primary aluminium time period situation with US electricity Aluminum, secondary, US-EI Unspecified Avg. European 97.1% from old scrap, at plant time period situation with US electricity Cardboard Avoided product US-EI Unspecified Avg. European - Corrugated board, fresh time period situation with bibre, single wall, at US electricity plant Corrugated board, US-EI Unspecified Avg. European 100% recycling fibre, single time period situation with -the mass of the input waste wall, at plant US electricity material is 97.1% of the mass of the product output Glass Avoided product: Adaptation of Unspecified Avg. European - Packaging glass, brown, US-EI based time period situation at plant/Virgin upon Magaud -Altered elect. et al. (2010) mix Packaging glass, brown, Adaptation of Unspecified Avg. European 100% at plant/100% Cullet US-EI based time period situation upon Magaud -Altered elect. et al. (2010) mix Avoided product: Adaptation of Unspecified Avg. European - Packaging glass, green, US-EI based time period situation at plant/Virgin upon Magaud -Altered elect. et al. (2010) mix Packaging glass, green, Adaptation of Unspecified Avg. European 100% at plant/100% Cullet US-EI based time period situation upon Magaud -Altered elect. et al. (2010) mix Avoided product: Adaptation of Unspecified Avg. European - Packaging glass, white, US-EI based time period situation at plant/Virgin upon Magaud -Altered elect. et al. (2010) mix Packaging glass, white, Adaptation of Unspecified Avg. European 100% at plant/100% Cullet US-EI based time period situation upon Magaud -Altered elect. et al. (2010) mix HDPE Avoided product US-EI Unspecified Avg. European - Polyethylene, HDPE, time period situation with granulate US electricity HDPE recycling1 Author- 2010 United States 92.7% defined unit process based upon Franklin Associates 2010 LDPE Avoided product US-EI Unspecified North America - Low density time period polyethylene resin, at plant NREL/RNA LDPE film recycled FAL Franklin USA 1995-1999 North America 91% 98 -the mass of the input waste material is 55.0% of the mass of the product output Paper Avoided product US-EI Unspecified Avg. European - Paper, newsprint, 0% time period situation with DIP, at plant US electricity Paper, newsprint, at US-EI Unspecified Avg. European 97.3% 285

plant2 time period situation with -the mass of the input waste US electricity material is 85.6% of the mass of the product output PET Avoided product US-EI Unspecified Avg. European - Polyethylene time period situation with terephthalate, granulate, US electricity bottle grade, at plant PET recycling Author- 2010 United States 80.0% defined unit process based upon Franklin Associates 2010 Polystyrene Avoided product Franklin USA 1995-1999 North America - (PS) PS (EPS) FAL 98 PS (EPS) recycled FAL Franklin USA 1995-1999 North America 91% 98 -the mass of the input waste material is 55.0% of the mass of the product output Steel Avoided product US-EI Unspecified Global average Pig iron, at plant time period with US electricity Iron scrap, at plant US-EI Unspecified Avg. European 100% time period situation with US electricity 1 In order to evaluate the production process for HDPE with 15% recycled content, as in WPA type 2, 16.2% of the HDPE production is depicted with the unit process: “HDPE recycling.” 2 In order to evaluate the production process for newsprint with 40% recycled content, as in WPA types 1 and 3, 48.0% of the newsprint production is depicted with the unit process: “Paper, newsprint, at plant.”

A4.5 Unit process inputs

For each scenario, Tables A4.8 and A4.9 list the LCA processes, represented by their codes, used in the calculation of their net environmental burdens, and specifies the quantities of each unit process required. The unique codes were supplied by the author to facilitate the identification of each unit process used in the WasteMAP LCA scenarios.

Table A4.8 Upstream unit processes and the quantities required under the waste prevention scenario

Process category Name of process Unit Waste prevention scenario Containers Glass bottles Packaging glass, brown, at plant Tonne -7.89*102 Packaging glass, green, at plant Tonne -3.91*103 Packaging glass, white, at plant Tonne -4.80*103 Aseptic cartons Aluminium, primary, at plant Tonne 7.62*10-1 Solid bleached board, SBB, at plant Tonne 1.40*101 Packaging film, LDPE, at plant Tonne 2.61*100 1 litre aseptic carton manufacture Tonne of aseptic 3.72*10-1 carton inputs 1.5 litre aseptic carton manufacture Tonne of aseptic 1.70*101 carton inputs PET bottles Polyethylene terephthalate, Tonne 4.93*101 granulate, bottle grade, at plant Blow molding Tonne of PET 4.93*101 inputs 286

Injection molding Tonne of PET 4.93*101 inputs Secondary packaging (capsules, closures and paper labels) Capsules Aluminium, primary, at plant Tonne -3.50*10-1 Packaging film, LDPE, at plant Tonne -3.10*10-1 Polyvinylchloride, at regional Tonne -1.10*10-1 storage (transport emissions removed) Tin, at regional storage Tonne -4.94*10-1 Closures Aluminium, primary, at plant Tonne -1.74*100 Kraft paper, bleached, at plant Tonne -8.52*10-3 Polyethylene terephthalate, Tonne granulate, amorphous, at plant -1.36*10-1 Polyethylene, HDPE, granulate, at Tonne plant 7.54*10-2 Polyethylene, LDPE, granulate, at Tonne plant -1.64*10-1 Polypropylene, granulate, at plant Tonne 7.54*10-2 Polyvinylchloride, at regional Tonne storage (transport emissions removed) -1.16*10-2 Raw cork, at forest road Tonne -4.12*10-1 Steel, cold rolled sheet, at plant Tonne NREL/RNA -1.06*100 Tin, at regional storage Tonne -3.88*10-2 Labels – paper (for glass Paper, wood-containing, LWC, at Tonne -6.20*10-1 and PET wine and spirits regional storage containers) Upstream transportation WPA type 1 Transportation of Transport, freight, rail, diesel Tonne-km -4.06*106 newsprint Transport, lorry >28t, fleet average Tonne-km -1.35*106 WPA type 2 Transportation of bags Transport, lorry >28t, fleet average Tonne-km -1.22*105 WPA type 3 Transportation of Transport, freight, rail, diesel Tonne-km -1.84*106 newsprint Transport, lorry >28t, fleet average Tonne-km -6.12*105 WPA types 4 and 5 Containers from Transport, freight, rail, diesel Tonne-km -3.89*106 manufacturer to Transport, lorry >28t, fleet average Tonne-km -3.89*106 winery/distillery Transport, transoceanic freight ship Tonne-km -5.77*106 Containers from Transport, freight, rail, diesel Tonne-km -4.65*106 winery/distillery to Transport, lorry >28t, fleet average Tonne-km -6.22*106 product distribution centre Transport, transoceanic freight ship Tonne-km -8.36*107

Table A4.9 Downstream unit processes and the quantities required under each WasteMAP scenario

Process category Name of process Unit 2008 reference Waste scenario prevention scenario Waste collection Transport, municipal waste Tonne-km 1.64*107 1.60*107 collection, lorry 21t Waste transport Transport, lorry >28t, fleet average Tonne-km 2.75*108 2.70*108 Transport, freight, rail Tonne-km 1.11*108 1.06*108 Transport, transoceanic freight ship Tonne-km 2.49*108 2.37*108

Sorting Recyclables, sorted at MRF, for Tonne 1.91*105 1.78*105 further treatment/Ontario 2008 Municipal recycling Aluminum Avoided product: Primary Tonne -1.06*103 -1.06*103 287

aluminium, at plant Aluminum, secondary, from old Tonne 1.06*103 1.06*103 scrap, at plant Cardboard/Polycoat Avoided product: Corrugated board, Tonne -3.26*104 -3.26*104 fresh fibre, single wall, at plant Corrugated board, recycling fibre Tonne 3.26*104 3.26*104 (single wall) Glass Avoided product: Packaging glass, Tonne -1.78*104 -1.66*104 green, at plant/Virgin Packaging glass, green, at Tonne 1.78*104 1.66*104 plant/100% Cullet HDPE Avoided product: Polyethylene, Tonne -2.20*103 -2.20*103 HDPE, granulate, at plant Process based upon Franklin Tonne 2.20*103 2.20*103 Associates 2010 LDPE Avoided product: Low density Tonne -1.73*101 -1.66*101 polyethylene resin, at plant NREL/RNA (US LCI database) LDPE film recycled FAL (Franklin Tonne 1.73*101 1.66*101 Associates 98) Paper Avoided product Tonne -1.01*105 -8.99*104 Paper, newsprint, 0% DIP, at plant Paper, newsprint, at plant Tonne 1.01*105 8.99*104 PET Avoided product: Polyethylene Tonne -2.42*103 -2.43*103 terephthalate, granulate, bottle grade Process based upon Franklin Tonne 2.42*103 2.43*103 Associates 2010 Polystyrene Avoided product: PS (EPS) FAL Tonne -7.45*101 -7.45*101 PS (EPS) recycled FAL Tonne 7.45*101 7.45*101 Steel/Other metals Avoided product: Pig iron, at plant Tonne -4.76*103 -4.76*103 Iron scrap, at plant Tonne 4.76*103 4.76*103 Recycling through deposit-return system Glass Avoided product: Packaging glass, Tonne -9.16*102 -3.82*102 brown, at plant/Virgin Packaging glass, brown, at Tonne 9.16*102 3.82*102 plant/100% Cullet Avoided product: Packaging glass, Tonne -6.27*103 -3.62*103 green, at plant/Virgin Packaging glass, green, at Tonne 6.27*103 3.62*103 plant/100% Cullet Avoided product: Packaging glass, Tonne -6.17*103 -2.93*103 white, at plant/Virgin Packaging glass, white, at Tonne 6.17*103 2.93*103 plant/100% Cullet Paper Avoided product Tonne -4.75*101 -3.18*101 Paper, newsprint, 0% DIP, at plant Paper, newsprint, at plant Tonne 4.75*101 3.18*101 Polycoat Avoided product: Corrugated board, Tonne -4.80*100 -9.49*100 fresh fibre, single wall, at plant Corrugated board, recycling fibre Tonne 4.80*100 9.49*100 (single wall) PET Avoided product: Polyethylene Tonne -3.04*101 -4.69*101 terephthalate, granulate, bottle grade Process based upon Franklin Tonne 3.04*101 4.69*101 Associates 2010 Biological treatment SSO and yard waste Compost, at plant Tonne 6.26*104 6.11*104 SSO anaerobic digestion Tonne 1.18*105 1.18*105 Bottle washing Bottle washing Bottle washing (author defined) Bottles 0 1.11*107

288

A4.6 Additional WasteMAP LCA results

Table A4.10 Midpoint level environmental impacts from the 2008 reference and waste prevention scenarios under the Impact 2002+ LCIA method.

Impact category Unit 2008 reference scenario Waste prevention scenario 6 6 Carcinogens kg C2H3Cl eq -8.15*10 -9.18*10 6 6 Non-carcinogens kg C2H3Cl eq -7.18*10 -7.54*10 4 4 Respiratory inorganics kg PM2.5 eq 6.08*10 2.92*10 Ionizing radiation Bq C-14 eq -8.83*109 -9.62*109 Ozone layer depletion kg CFC-11 eq 7.71*100 4.88*100 4 4 Respiratory organics kg C2H4 eq 6.36*10 5.19*10 Aquatic ecotoxicity kg TEG water 2.22*1010 2.05*1010 Terrestrial ecotoxicity kg TEG soil 6.65*108 2.32*108 6 6 Terrestrial acid/nutri kg SO2 eq 2.92*10 2.20*10 Land occupation m²org.arable -5.69*107 -5.65*107 5 4 Aquatic acidification kg SO2 eq 2.81*10 8.79*10 3 3 Aquatic eutrophication kg PO4 P-lim 6.20*10 2.55*10 7 6 Global warming kg CO2 eq 2.41*10 -5.14*10 Non-renewable energy MJ primary -9.94*108 -1.54*109 Mineral extraction MJ surplus -3.32*106 -3.96*106

Table A4.11 Midpoint level environmental impacts from the 2008 reference and waste prevention scenarios under the ReCiPe (H) LCIA method.

Impact category Unit 2008 reference scenario Waste prevention scenario 8 8 Climate change kg CO2 eq 2.95*10 2.61*10 Ozone depletion kg CFC-11 eq 8.05*100 5.23*100 5 4 Terrestrial acidification kg SO2 eq 1.23*10 -5.03*10 Freshwater eutrophication kg P eq 4.84*103 3.62*103\ Marine eutrophication kg N eq 2.59*106 2.53*106 Human toxicity kg 1,4-DB eq 2.44*108 2.25*108 Photochemical oxidant kg NMVOC 6.19*105 4.82*105 formation 4 3 Particulate matter formation kg PM10 eq 6.17*10 2.47*10 Terrestrial ecotoxicity kg 1,4-DB eq 1.84*104 1.17*104 Freshwater ecotoxicity kg 1,4-DB eq 1.03*107 1.00*107 Marine ecotoxicity kg 1,4-DB eq 9.47*106 9.21*106 Ionising radiation kg U235 eq -8.81*107 -9.59*107 Agricultural land occupation m²a -5.20*108 -5.13*108 Urban land occupation m²a -6.30*106 -6.35*106 Natural land transformation m² -4.44*104 -5.06*104 Water depletion m³ -8.35*105 -1.24*106 Metal depletion kg Fe eq -5.96*106 -7.55*106 Fossil depletion kg oil eq 7.49*106 -2.48*106

Table A4.12 Midpoint level environmental impacts from the 2008 reference and waste prevention scenarios under the TRACI 2 LCIA method.

Impact category Unit 2008 reference scenario Waste prevention scenario 8 8 Global warming kg CO2 eq 6.16*10 5.80*10 Acidification H+ moles eq 9.44*106 -6.45*105 Carcinogenics kg benzene eq 6.15*107 6.13*107 Non carcinogenics kg toluene eq 2.01*1012 2.00*1012 4 5 Respiratory effects kg PM2.5 eq -6.74*10 -1.13*10 Eutrophication kg N eq 3.69*106 3.61*106 Ozone depletion kg CFC-11 eq 7.78*100 4.96*100 Ecotoxicity kg 2,4-D eq 1.96*1010 1.94*1010 5 5 Smog kg NOx eq 4.56*10 3.48*10 289

Table A4.13 Endpoint level environmental impacts from the 2008 reference and waste prevention scenarios under the Impact 2002+ LCIA method.

Impact category Unit 2008 reference scenario Waste prevention scenario Carcinogens DALY -2.28*101 -2.57*101 Non-carcinogens DALY -2.01*101 -2.11*101 Respiratory inorganics DALY 4.26*101 2.05*101 Ionizing radiation DALY -1.86*100 -2.02*100 Ozone layer depletion DALY 8.10*10-3 5.13*10-3 Respiratory organics DALY 1.36*10-1 1.11*10-1 Aquatic ecotoxicity PDF*m²*yr 1.12*106 1.03*106 Terrestrial ecotoxicity PDF*m²*yr 5.26*106 1.83*106 Terrestrial acid/nutri PDF*m²*yr 3.04*106 2.29*106 Land occupation PDF*m²*yr -6.20*107 -6.16*107 7 6 Global warming kg CO2 eq 2.41*10 -5.14*10 Non-renewable energy MJ primary energy -9.94*108 -1.54*109 Mineral extraction MJ primary energy -3.32*106 -3.96*106

Table A4.14 Endpoint level environmental impacts from the 2008 reference and waste prevention scenarios under the ReCiPe (H) LCIA method.

Impact category Unit 2008 reference scenario Waste prevention scenario Climate change – human DALY 4.18*102 3.70*102 health Climate change - ecosystems species*yr 2.36*100 2.09*100 Ozone depletion DALY 2.55*10-2 1.84*10-2 Terrestrial acidification species*yr 7.17*10-4 -2.91*10-4 Freshwater eutrophication species*yr 2.15*10-4 1.61*10-4 Human toxicity DALY 1.71*102 1.57*102 Photochemical oxidant DALY 2.42*10-2 1.88*10-2 formation Particulate matter formation DALY 1.60*101 6.42*10-1 Terrestrial ecotoxicity species*yr 2.34*10-3 1.49*10-3 Freshwater ecotoxicity species*yr 2.67*10-3 2.60*10-3 Marine ecotoxicity species*yr 7.58*10-6 7.37*10-6 Ionising radiation DALY -1.44*100 -1.57*100 Agricultural land occupation species*yr -5.81*100 -5.74*100 Urban land occupation species*yr -1.22*10-1 -1.23*10-1 Natural land transformation species*yr 1.39*10-1 8.13*10-2 Metal depletion $ -4.26*105 -5.40*105 Fossil depletion $ 1.20*10\8 -4.02*107

Table A4.15 Identification of the most important process contributors to impacts and avoided burdens for each midpoint level category

Midpoint impact Impact / Process category Avoided burden Impact 2002+ Carcinogens Impact Hydrogen peroxide, 50% in H2O, at plant Avoided burden Natural gas, at production Non-carcinogens Impact Disposal, MSW, 22.9% water, to sanitary landfill Avoided burden Paper, newsprint, 0% deinked paper, at plant Respiratory Impact Operation, lorry >28t, fleet average inorganics Avoided burden Packaging glass, green, at plant/virgin Ionizing radiation Impact Titanium dioxide, chloride process, at plant Avoided burden Tailings, uranium milling Ozone layer Impact Crude oil, at production onshore Depletion Avoided burden Uranium, enriched 3.8%, at USEC enrichment plant1 290

Respiratory Impact Transport, municipal waste collection, lorry 21t Organics Avoided burden Power sawing, without catalytic converter Aquatic Impact Disposal, aluminum, 0% water, to sanitary landfill ecotoxicity Avoided burden Uranium natural, in yellowcake, at mill plant Terrestrial Impact Operation, lorry >28t, fleet average ecotoxicity Avoided burden Blasting Terrestrial Impact Operation, lorry >28t, fleet average acidification / Avoided burden Ref: Packaging glass, green, at plant/virgin nutrification WP: Paper, newsprint, 0% deinked paper, at plant Land occupation Impact Sanitary landfill facility Avoided burden Softwood, Scandinavian, standing, under bark, in forest Aquatic Impact Compost, at plant acidification Avoided burden Packaging glass, green, at plant/virgin Aquatic Impact Paper, newsprint, at plant eutrophication Avoided burden Paper, newsprint, 0% deinked paper, at plant Global warming Impact Paper, newsprint, at plant Avoided burden Ref: Packaging glass, green, at plant/virgin WP: Hard coal, burned in power plant Non-renewable Impact Crude oil, at production onshore energy Avoided burden Uranium natural, at underground mine Mineral extraction Impact Ferronickel, 25% Ni, at plant Avoided burden Bauxite, at mine ReCiPe Climate change Impact Disposal, MSW, 22.9% water, to sanitary landfill Avoided burden Ref: Packaging glass, green, at plant/virgin WP: Hard coal, burned in power plant Ozone depletion Impact Crude oil, at production onshore Avoided burden Uranium, enriched 3.8%, USEC enrichment plant1 Terrestrial Impact Compost, at plant Acidification Avoided burden Packaging glass, green, at plant/virgin Freshwater Impact Disposal, MSW, 22.9% water, to sanitary landfill eutrophication Avoided burden Disposal, ash from paper production, sludge, 0% water, to residual waste landfill Marine Impact Disposal, MSW, 22.9% water, to sanitary landfill eutrophication Avoided burden Paper, newsprint, 0% deinked paper, at plant Human toxicity Impact Disposal, MSW, 22.9% water, to sanitary landfill Avoided burden Paper, newsprint, 0% deinked paper, at plant Photochemical Impact Operation, lorry >28t, fleet average oxidant formation Avoided burden Paper, newsprint, 0% DIP, at plant Particulate matter Impact Operation, lorry >28t, fleet average formation Avoided burden Packaging glass, green, at plant/virgin Terrestrial Impact Palm fruit bunches, at farm ecotoxicity Avoided burden Packaging glass, green, at plant/virgin Freshwater Impact Disposal, MSW, 22.9% water, to sanitary landfill Ecotoxicity Avoided burden Ref: Paper, newsprint, 0% deinked paper, at plant WP: Natural gas, unprocessed, at extraction Marine ecotoxicity Impact Disposal, MSW, 22.9% water, to sanitary landfill Avoided burden Disposal, redmud from bauxite digestion, 0% water, to residual material landfill Ionising radiation Impact Titanium dioxide, chloride process, at plant Avoided burden Tailings, uranium milling Agricultural land Impact Green manure IP, unit March Occupation Avoided burden Softwood, Scandinavian, standing, under bark, in forest Urban land Impact Sanitary landfill facility Occupation Avoided burden Industrial wood, Scandinavian softwood, under bark, u=140%, at forest road Natural land Impact Well for exploration and production, onshore transformation Avoided burden Industrial wood, Scandinavian softwood, under bark, u=140%, at forest road Water depletion Impact Paper, newsprint, at plant Avoided burden Paper, newsprint, 0% deinked paper, at plant Metal depletion Impact Ref: Molybdenum concentrate, main product WP: Zinc concentrate, at benefication Avoided burden Iron ore, 46% FE, at mine Fossil fuel Impact Crude oil, at production onshore Depletion Avoided burden Hard coal, at mine TRACI 2 291

Global warming Impact Disposal, MSW, 22.9% water, to sanitary landfill Avoided burden Paper, newsprint, 0% deinked paper, at plant Acidification Impact Operation, lorry >28t, fleet average Avoided burden Packaging glass, green, at plant Carcinogenics Impact Disposal, MSW, 22.9% water, to sanitary landfill Avoided burden Paper, newsprint, 0% deinked paper, at plant Non-carcinogenics Impact Disposal, MSW, 22.9% water, to sanitary landfill Avoided burden Ref: Soda, powder, at plant Paper, newsprint, 0% deinked paper, at plant Respiratory effects Impact Paper, newsprint, at plant Avoided burden Packaging glass, green, at plant Eutrophication Impact Disposal, MSW, 22.9% water, to sanitary landfill Avoided burden Paper, newsprint, 0% deinked paper, at plant Ozone depletion Impact Crude oil, at production onshore Avoided burden Uranium, enriched 3.8%, at USEC enrichment plant1 Ecotoxicity Impact Disposal, MSW, 22.9% water, to sanitary landfill Avoided burden Disposal, redmud from bauxite digestion, 0% water, to residual material landfill Smog Impact Operation, lorry >28t, fleet average Avoided burden Paper, newsprint, 0% deinked paper, at plant 1 As there is no published unit process for Canadian heavy water nuclear reactors, the unit process used to depict nuclear energy addresses the conditions for generic light water reactors, which require enrichment of the uranium fuel. Uranium enrichment is not necessary for Canadian heavy water reactors.

A4.7 References

City of Toronto. 2006. Backgrounder: Release of 2006 Census results: Population and Dwelling Counts. Available at: http://www.toronto.ca/demographics/pdf/ 2006_population_and_dwelling_count_backgrounder.pdf. Accessed on 2010 09 15.

Fisher, C. 2008. 2008 Capsule Report: Small Wineries Maintain Course with Tin Capsules Despite Skyrocketing Costs. Wine Business Monthly. September 15, 2008. Available at: http://www.winebusiness.com/wbm/?go=getArticle&dataId=59224. Accessed on 2010 06 21.

Google. 2010. Google Maps. Available at: http://maps.google.ca/maps. Accessed on 2010 02 18.

Stewardship Ontario. 2006b. Single Family Waste Audit Program. Toronto Single Family Dwellings: Winter 2005, Spring 2005, Summer 2005, Fall 2005. Available at: 292

http://www.stewardshipontario.ca/eefund/projects/audits/waste_audit_sf.htm. Accessed on 2009 03 16.

Tetra Pak. 2005. The Recycling Brochure for Europe. Available at: http://www.tetrapak.com/docs/recycling_booklet.pdf. Accessed on 2008 01 20.

Toronto Star n.d. Frequently asked questions. Available at: http://www.thestar.com/faq#printing. Accessed on: 2011 02 02.

Travacon Research Limited. 2007. Final report: An estimate of the freight rate consequences of rail captivity to rail shippers of Canadian forest products. Prepared for Forest Products Association of Canada. Available at: http://www.fpac.ca/publications/Travacon_final_report.pdf. Accessed on 2011 04 13.

World-Ports Distances Calculator. 2010. Available at http://www.distances.com/. Accessed on 2010 08 07.