Text Material Preview
Life Cycle Assessment Handbook Scrivener Publishing 100 Cummings Center, Suite 541J Beverly, MA 01915-6106 Publishers at Scrivener Martin Scrivener (martin@scrivenerpublishing.com) Phillip Carmical (pcarmical@scrivenerpublishing.com) Life Cycle Assessment Handbook A Guide for Environmentally Sustainable Products Edited by Mary Ann Curran Cincinnati, OH, USA Scrivener WILEY Copyright © 2012 by Scrivener Publishing LLC. All rights reserved. Co-published by John Wiley & Sons, Inc. Hoboken, New Jersey, and Scrivener Publishing LLC, Salem, Massachusetts. Published simultaneously in Canada. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., I l l River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/ permission. Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the con- tents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com. For more information about Scrivener products please visit www.scrivenerpublishing.com. Cover design by Kris Hackerott. Library of Congress Cataloging-in-Publication Data: ISBN 978-1-118-09972-8 Printed in the United States of America 10 9 8 7 6 5 4 3 2 1 Contents Preface xix 1 Environmental Life Cycle Assessment: Background and Perspective 1 Gjalt Huppes and Mary Ann Curran 1.1 Historical Roots of Life Cycle Assessment 1 1.2 Environmental Life Cycle Concepts 2 1.3 LCA Links to Environmental Policy 3 1.4 Micro Applications of LCA Rising 5 1.5 The Micro-Macro Divide 5 1.6 Macro Level LCA for Policy Support 6 1.7 Example Biofuels 7 1.8 Why Environmental LCA? 8 1.9 Overview of the Book 11 1.9.1 Methodology and Current State of LCA Practice 11 1.9.2 LCA Applications 12 1.9.3 LCA Supports Decision Making and Sustainability 13 1.9.4 Operationalizing LCA 13 References 14 Part 1: Methodology and Current State of LCA Practice 2 An Overview of the Life Cycle Assessment Method - Past, Present, and Future 15 Reinout Heijungs and Jeroen B. Guinee 2.1 The Present-Day LCA Method 15 2.1.1 Goal and Scope Definition 17 2.1.2 Inventory Analysis 18 2.1.3 Impact Assessment 22 2.1.4 Interpretation 27 2.1.5 LCA in Practice 29 2.2 A Short History of LCA 30 2.2.1 Past LCA (1970-2000): Conception and Standardization 30 2.2.1.1 1970-1990: Decades of Conception 30 2.2.1.2 1990-2000: Decade of Standardization 31 2.2.2 Present LCA (2000-2010): Decade of Elaboration 32 2.2.3 Future LCA (2010-2020): Decade of Life Cycle Sustainability Analysis 34 References 37 vi CONTENTS 3. Life Cycle Inventory Modeling in Practice 43 Beverly Sauer 3.1 Introduction 43 3.2 Study Goal 44 3.3 Scope 45 3.3.1 Functional Unit 45 3.3.2 Boundaries 47 3.4 Methodology Issues 55 3.4.1 Feedstock Energy 55 3.4.2 Multi-Output Processes 57 3.4.3 Postconsumer Recycling 58 3.4.4 Converting Scrap 60 3.4.5 Water Use 61 3.4.6 Carbon Tracking Considerations 62 3.5 Evolution of LCA Practice and Associated Issues 63 3.6 Conclusion 65 References 65 4 Life Cycle Impact Assessment 67 Manuele Margni and Mary Ann Curran 4.1 Introduction 67 4.2 Life Cycle Impact Assessment According to ISO 14040-44 Requirements 69 4.2.1 Overview 69 4.2.2 Mandatory Elements 70 4.2.3 Optional Elements 72 4.2.4 Interpreting an LCIA Profile 73 4.3 Principles and Framework of LCIA 74 4.4 Historical Developments and Overview of LCIA Methodologies 78 4.5 Variability in the LCIA Models 86 4.6 State-of-the-Art LCIA 88 4.7 Future Development 94 4.7.1 Spatially-Differentiated Assessment in LCIA 94 4.7.2 Addressing Uncertainty and Variability in Characterization Factors 95 4.7.3 Improving the Characterization of Resources 96 4.7.4 Integrating Water Use and Consumption in LCIA 97 4.7.5 Resources and Ecosystem Services Areas of Protection 98 4.7.6 Expanding Land Use Burdens on Biodiversity in Ecosystem Services 99 References 99 5 Sourcing Life Cycle Inventory Data 105 Mary Ann Curran 5.1 Introduction 105 5.2 Developing LCI to Meet the Goal of the Study 107 5.2.1 Considerations in Choosing Data Sources 107 5.2.2 A Word on Consequential Life Cycle Assessment 108 CONTENTS vii 5.3 Types of LCI Data 109 5.4 Private Industrial Data 112 5.5 Public Industrial Data 112 5.6 Dedicated LCI databases 113 5.7 Using Non-LCI Data in LCAs 118 5.8 Creating Life Cycle Inventory using Economic Input/Output Data 134 5.9 Global Guidance for Database Creation and Management 135 5.10 Future Knowledge Management 136 5.10.1 Creating a Federal Data Commons in the US 137 5.10.2 Open-Source Models 138 5.10.3 Crowdsourcing 139 5.11 Conclusion 140 References 141 6 Software for Life Cycle Assessment 143 Andreas Ctroth 6.1 LCA and LCA Software 143 6.1.1 Introduction 143 6.1.2 Characteristics of LCA Software Systems 144 6.1.2.1 Web Tools versus Desktop Tools 144 6.1.2.2 Commercial Tools versus Freeware 145 6.1.2.3 Open Source versus Closed Source 146 6.1.2.4 General LCA Tools versus Specialised Tools versus Add-Ons 147 6.1.3 Two Basic LCA Software User Types and their Needs 149 6.1.4 The LCA Software Market 150 6.1.4.1 Main LCA Software Systems 150 6.1.4.2 Other LCA Software Systems 152 6.1.5 Trends in LCA Software 152 6.1.5.1 Ideas that are No Longer Trends 153 6.1.5.2 Possible Future Trends 155 6.1.6 Outlook and Conclusions 156 References 157 Part 2: LCA Applications 7 Modeling the Agri-Food Industry with Life Cycle Assessment 159 Bruno Notarnicola, Giuseppe Tassielli and Pietro A. Renzulli 7.1 Introduction 159 7.2 Methodological Issues 161 7.2.1 Choice of Functional Unit 161 7.2.2 System Boundaries, Carbon Balance and Data Quality 165 7.2.3 Fertilizer and Pesticide Dispersion Models 167 7.2.4 Land Use and Water Use Impact Categories 170 7.2.4.1 Land Use 170 7.2.4.2 Water Use 173 7.3 Role of the Food Industry: Some Examples 174 7.4 Conclusions 177 References 178 CONTENTS Exergy Analysis and its Connection to Life Cycle Assessment 185 Marc A. Rosen, Ibrahim Dincer and Ahmet Ozbilen 8.1 Introduction 185 8.2 Life Cycle Assessment 187 8.2.1 Goal and Scope Definition 188 8.2.2 Life Cycle Inventory Analysis 188 8.2.3 Life Cycle Impact Assessment 188 8.2.4 Life Cycle Interpretation (Improvement Analysis)190 8.3 Exergy and Exergy Analysis 190 8.3.1 Characteristics of Exergy 190 8.3.2 Exergy Analysis 191 8.4 Exergetic Life Cycle Assessment (ExLCA) 192 8.4.1 Linkages between Exergy Analysis and LCA 192 8.4.2 Rationale of ExLCA 194 8.4.3 ExLCA Methodology and Approach 195 8.4.4 Applications of ExLCA 196 8.4.5 Advantages of ExLCA 199 8.5 Case Study 199 8.5.1 System Description and Data Analysis 201 8.5.1.1 Hydrogen Production Plant Based on a Cu-Cl Thermochemical Cycle 202 8.5.1.2 Nuclear Plant 204 8.5.1.3 Fuel (Uranium) Processing 204 8.5.2 Analysis 205 8.5.2.1 LCA of Overall System 205 8.5.2.2 ExLCA of Overall System 206 8.5.3 LCA and ExLCA Results and Discussion 208 8.6 Conclusions 211 Acknowledgements 212 Nomenclature 212 Acronyms 212 References 213 Accounting for Ecosystem Goods and Services in Life Cycle Assessment and Process Design 217 Erin F. Landers, Robert A. Urban and Bhavtk R. Baksht 9.1 Motivation 217 9.2 Life Cycle Assessment Background 219 9.3 Ecologically-Based Life Cycle Assessment 220 9.4 Case Study Comparing Process-Based and Hybrid Studies Based on EIO-LCA and Eco-LCA 222 9.5 Overview of the Role of Ecosystems in Sustainable Design 226 9.6 Design Case Study: Integrated Design of a Residential System 227 9.7 Conclusions 229 References 230 CONTENTS ix A Case Study of the Practice of Sustainable Supply Chain Management 233 Annie Wetsbrod and Larry Loftus 10.1 Introduction 233 10.2 Why Develop an Integrated Sustainable Supply Chain Management Program? 235 10.3 How Might the World's Largest Consumer Products Company Measure and Drive Sustainability in its Supply Chains? 238 10.4 What is the State of P&G's Supply Chain Environmental Sustainability? 240 10.5 Why is the Scorecard Effective for Driving Change and Building Environmental Tracking Capability? 245 10.6 What is involved with Social Sustainability in Supply Chain Management? 247 10.7 Conclusion 248 References 248 Life Cycle Assessment and End of Life Materials Management 249 Keith A. Weitz 11.1 Introduction 249 11.2 Value of Applying Life Cycle Principles and Concepts to End-Of-Life Materials Management 250 11.3 LCA of Waste Management Versus GHG Inventory/Reporting, Sustainability Reporting, and Other Environmental Initiatives 251 11.4 Summary of Key Life Cycle Procedures and their Application to End-Of-Life Systems 255 11.4.1 Goals and Scope 256 11.4.2 System Function and Functional Unit 256 11.4.3 Boundary Decisions 256 11.4.4 Geographic Boundaries 259 11.4.5 Time Scale Boundaries 260 11.4.6 Key LCA Modeling Decision Points 260 11.5 Overview of Existing Waste Related LCAs 261 11.6 Using Waste Management LCA Information for Decision Making 265 References 265 Application of LCA in Mining and Minerals Processing - Current Programs and Noticeable Gaps 267 Dr. Mary Stewart, Dr. Peter Holt and Mr. Rob Rouwette 12.1 Introduction 267 12.2 The Status Quo 268 12.2.1 LCA Use in the Mining and Mineral Processing Industry 268 12.2.1.1 Low Overall Business Priority 271 CONTENTS 12.2.2 Life Cycle Inventory/Life Cycle Assessment in Mining and Processing 272 12.2.2.1 Corporate Initiatives 272 12.2.2.2 Association Initiatives 273 12.2.2.3 Supply Chain and Voluntary Initiatives 274 12.2.2.4 Market Positioning and Advocacy 276 12.2.3 Life Cycle Management 276 12.3 What is LCA and LCM Information Being Used for? 279 12.3.1 Internal Decision Taking 280 12.3.2 External Decision Taking 281 12.4 Gaps and Constraints 284 12.4.1 Methodological Considerations 284 12.4.2 Value Chain Structures 286 12.5 Conclusions and Recommendations 288 References 289 Sustainable Preservative-Treated Forest Products, Their Life Cycle Environmental Impacts, and End of Life Management Opportunities: A Case Study 291 Christopher A. Bolin 13.1 Introduction 291 13.2 Life Cycle Inventory Analysis 293 13.2.1 Forestry and Milling 293 13.2.1.1 Forestry 293 13.2.1.2 Milling 294 13.2.1.3 Properties of Wood 295 13.2.2 Preservative Manufacture and Treatment of Lumber Products 296 13.2.3 Preservative-Treated Wood Product Service Life 299 13.2.4 End of Life Management 299 13.2.4.1 Landfill Disposal 299 13.2.4.2 Reuse 300 13.2.4.3 Reuse for Energy 300 13.3 Energy Reuse Considerations 301 13.3.1 Chemicals in Preservative-treated Wood 301 13.3.1.1 Lumber Containing Copper-Based Preservative 301 13.3.1.2 Lumber Containing Boron-Based Preservatives 301 13.3.2 Lumber Collection at the End of Service Life 302 13.4 Case Study Scenarios 302 13.5 Carbon Accounting, Impact Indicator Definition, and Classification 303 13.5.1 Carbon Accounting 303 13.5.2 Fossil Fuel Usage 304 13.5.3 Total Energy 304 13.5.4 Other Impact Indicators Assessed 305 CONTENTS xi 13.6 Lumber Life Cycle Assessment Findings 305 13.7 Conclusions 308 References 308 Buildings, Systems Thinking, and Life Cycle Assessment 311 Joel Ann Todd 14.1 Introduction 311 14.2 Applying LCA to Buildings 314 14.2.1 Opportunities 314 14.2.2 Challenges 315 14.3 History and Progress in Applying LCA to Buildings 319 14.3.1 Databases, Tools, and Resources 319 14.3.1.1 AIA Environmental Resource Guide 319 14.3.1.2 BEES 320 14.3.1.3 US LCI Database 321 14.3.1.4 ATHENA Ecocalculator and Impact Estimator 321 14.3.1.5 Other Tools 321 14.3.2 International Standards and Codes 322 14.3.2.1 ISO 322 14.3.2.2 CENTC350 322 14.3.2.3 ANSI/ASHRAE/USGBC/IES Standard 189 for the Design of High-Perf ormance Green Buildings, Except Low-Rise Residential Buildings 323 14.3.2.4 International Green Construction Code (IGCC) 323 14.3.3 Assessment and Certification Systems 324 14.3.3.1 BREEAM 324 14.3.3.2 LEED 325 14.3.3.3 DGNB 325 14.3.3.4 Green Globes (US) 326 14.4 Evolution and Future Applications to the Built Environment 326 References 327 Life Cycle Assessment in Product Innovation 329 Nuno Da Silva 15.1 Introduction 329 15.2 Background 330 15.3 What R&D is For 331 15.4 The Innovation Funnel 331 15.5 Idea Generation 332 15.6 Idea Assessment 334 15.7 Concept Development 335 15.8 Business Planning and Execution 337 15.9 Where to Focus - Management Framework 337 15.10 Sustainable Portfolio Management 338 15.11 Tools 340 15.12 Data 342 References 342 CONTENTS Life Cycle Assessment as a Tool in Food Waste Reduction and Packaging Optimization - Packaging Innovation and Optimization in a Life Cycle Perspective 345 Ole Jörgen Haussen, Hanne Moller, Erik Svanes and Vibeke Schakenda 16.1 Introduction 345 16.2 Food Waste and Packaging Optimization in a Life Cycle Perspective 346 16.3 Principles and Models for Optimal Packaging in a Life Cycle/Value Chain Perspective 350 16.4 Case Studies on LCA of Food Waste and Packaging Optimization 354 16.4.1 Case Studies on Packaging Optimization and Food Waste Reduction? 354 16.4.2 Case Study on Coffee Packing and Distribution 355 16.4.2.1 Packaging System and Effects of Implemented Improvement Options 355 16.4.2.2 Effects of 20% Improvement in Strategies for Packaging Optimization 356 16.4.3 Case study on Packing and Distribution of Whole Pieces of Cheese 356 16.4.3.1 Optimization of Degree of Filling on Pallet for Cheese Packaging 357 16.4.3.2 Effects of 20% Improvement in Strategies for Packaging Optimization 358 16.4.3.3 Comparison of the Value Chain for Whole Pieces of Cheese and Sliced Cheese and the Corresponding Packaging 359 16.4.3.4 Effects of 20% Improvement in Strategies for Packaging Optimization 360 16.4.4 Case Study on Salad Packing and Distribution 361 16.5 Discussion and Conclusions 363 References 366 Integration of LCA and Life-Cycle Thinking within the Themes of Sustainable Chemistry & Engineering 369 Shawn Hunter, Richard Helling and Dawn Shiang 17.1 Introduction 369 17.2 The Four Themes of Sustainable Chemistry & Engineering 370 17.3 Life Cycle Assessment as a Tool for Evaluating SC&E Opportunities 376 17.3.1 Importance of LifeCycle Thinking for SC&E 376 17.3.2 What is the Value of a Renewable Feedstock? 378 17.3.2.1 Natural Oil-Based Polyols 378 17.3.2.2 Sugarcane-Based Polyethylene 380 17.3.3 How Important is the Project Team's Piece of the Life Cycle? 381 CONTENTS xiii 17.3.3.1 New Coatings Technology 382 17.3.3.2 LCA of Tetrahydrofuran Synthesis in High-Temperature Water 383 17.3.4 What is the Return on Life Cycle Investment? 384 17.4 LCA - One Tool in the Sustainability Toolbox 385 17.4.1 Screening Sustainability Assessment Tools 385 17.4.2 Economic Evaluation 386 17.4.3 Site-Specific Assessment Tools 386 17.4.3.1 Environmental Impact Assessment 387 17.4.3.2 Risk Assessment 387 17.4.3.3 Social Impact Assessment 387 17.5 Summary 388 Acknowledgement 388 References 388 Part 3: LCA Supports Decision Making and Sustainability 18 How to Approach the Assessment? 391 Jose Potting, Shabbtr Gheewala, Sebastten Bonnet and Joost van Buuren 18.1 Introduction 391 18.2 Assessment Methods 393 18.2.1 Technology Assessment 393 18.2.2 Environmental Impact Assessment 394 18.2.3 Risk Assessment 396 18.2.4 Life Cycle Assessment 398 18.3 Comparison of Assessment Methods 400 18.4 Guidance for Assessment 405 18.5 Discussion and Conclusions 409 Acknowledgement 410 References 410 19 Integration of MCDA Tools in Valuation of Comparative Life Cycle Assessment 413 Valenttna Prado, Kristen Rogers and Thomas P. Seager PhD 19.1 Introduction 413 19.2 Current Practices in LCIA 415 19.3 Principles of External Normalization 416 19.4 Issues with External Normalization 417 19.4.1 Inherent Data Gaps 417 19.4.2 Masking Salient Aspects 417 19.4.3 Compensation 419 19.4.4 Spatial Boundaries and Time Frames 419 19.4.5 Divergence in Data Bases 419 19.5 Principles of Internal Normalization 419 19.5.1 Compensatory Methods 420 19.5.2 Partially Compensatory Methods 421 CONTENTS 19.6 Weighting 423 19.7 Case 1: Magnitude Sensitivity 424 19.8 Case 2: Rank Reversal 426 19.9 Conclusions 428 References 428 Social Life Cycle Assessment: A Technique Providing a New Wealth of Information to Inform Sustainability-Related Decision Making 433 Catherine Benott Norris 20.1 Historical Development 433 20.2 Why Do Businesses Care? 435 20.3 Methodology 436 20.3.1 Defining Social Issues 436 20.3.2 The Framework 437 20.3.3 Typical Phases of a Study 441 20.3.3.1 Iterative Process of Social Life Cycle Assessment 441 20.3.3.2 Goal and Scope 442 20.3.3.3 Life Cycle Inventory 444 20.3.3.4 Life Cycle Impact Assessment 444 20.3.3.5 Interpretation 445 20.4 SLCA and other Key Social Responsibility References and Instruments 445 20.5 Conclusion 449 References 450 Life Cycle Sustainability Analysis 453 Alessandra Zamagni, Jeroen Guinee, Reinout Hetjungs and Paolo Masoni 21.1 LCA and Sustainability Questions 453 21.1.1 What is Sustainability? 453 21.1.2 Life Cycle Analysis and Sustainability 455 21.2 A Framework for Life Cycle Sustainability Analysis 459 21.2.1 Broadening 461 21.2.1.1 Broadening of the Object of Analysis 461 21.2.1.2 Broadening of the Spectrum of Indicators 462 21.2.2 Deepening 466 21.2.2.1 Increasing Sophistication in LCI Modelling 466 21.2.2.2 Economic and Behavioral Mechanisms 467 21.2.2.3 Deepening LCA and Consequential LCA 468 21.3 Future Directions for Research 469 21.3.1 Aligning Environmental with Economic and Social Indicators 470 21.3.2 Framing the Question 471 21.3.3 Modelling Options for Meso-Level and Economy-Wide Applications 471 References 472 CONTENTS XV 22 Environmental Product Claims and Life Cycle Assessment 475 Martha J. Stevenson and Wesley W. Ingwersen 22.1 Introduction 475 22.2 Typology of Claims: Three Different Claims per ISO Standards 477 22.2.1 Type I Ecolabels 477 22.2.2 Type II Environmental Claims 478 22.2.3 Type III Environmental Product Declarations 479 22.2.3.1 An EPD is a Document 479 22.2.3.2 An EPD is Primarily Based on LCA 479 22.2.3.3 An EPD is Developed by Following a "Product Category Rule" 480 22.2.3.4 An EPD can Contain Information Beyond the Scope of an LCA, Where Relevant to that Product 480 22.2.4 Further Information on EPDs and PCRs 481 22.2.5 Reference Case Study on Dairy PCR & EPDs 481 22.2.5.1 Liquid Milk PCR 482 22.2.5.2 Granarolo Milk EPD 483 22.3 Other LCA-Based Product Claims 484 22.4 Other Relevant Environmental Information 485 22.4.1 Water Footprinting 486 22.4.2 Toxicity Risk Assessment 486 22.4.3 Ecosystem Services Assessment 487 22.5 Conclusion 487 References 488 Appendix 1: Global Update of PCR/EPD Activity 491 Appendix 2: Product Category Rules 497 Appendix 3: Environmental Product Declaration for High-Quality Pasteurized Milk Packaged in Pet Bottles 521 Part 4: Operationalizing LCA 23 Building Capacity for Life Cycle Assessment in Developing Countries 545 Prof. Toolseeram Ramjeawon 23.1 Introduction 545 23.2 Status of LCA in Developing Countries 546 23.3 Challenges and opportunities 547 23.3.1 Challenges 547 23.3.2 Opportunities 549 23.4 Improving the Effectiveness of Capacity Building Initiatives 550 23.5 A Roadmap for Capacity Building in LCA in Developing Countries 555 23.5.1 Introduction of Life Cycle Topics in Educational Programs and Research Activities 556 23.5.2 Networking 558 CONTENTS 23.5.3 Setting up of a National Inventory Database and Development of Tools to Set Up, Maintain and Disseminate Data 558 23.5.4 Development of National Life Cycle Impact Assessment (LCIA) Methodologies 559 23.5.5 Capacity Development to Apply LCA in Industry and in Public Decision Making 559 23.5.6 Promotion of LCA Applications and Creating a Stock of Success Stories and Dissemination 560 23.5.7 Policy Development 560 23.6 Conclusions 560 References 561 Environmental Accountability: A New Paradigm for World Trade is Emerging 563 Ann K. Ngo 24.1 Introduction 563 24.2 The Paradigm Shift and LCA 564 24.3 International Trade and LCA 568 24.4 Behavior Change and LCA 570 24.4.1 The Role of Businesses 571 24.4.2 The Role of Governments 572 24.4.3 The Role of Consumers 576 24.4.4 The Role of NGOs 577 24.4.5 The Role of Academia 578 24.5 Challenges and Opportunities for a World Shifting to Using LCA and Environmental Impacts as Components of Regulation and Commerce 580 Appendix I 582 References 583 Life Cycle Knowledge Informs Greener Products 585 James Fava 25.1 Introduction 585 25.2 Situation Analysis 586 25.2.1 How Could We Set a River on Fire? 586 25.2.2 After an Early LCA Study, Coca-Cola Opted to Challenge its Suppliers to Improve their Products Rather than Simply Prohibiting the Use of Certain Materials 587 25.2.3 Dueling Diaper Debates Fueled the Initial Understanding that all Products have Impacts that may differ in Nature, Scope, and Medium 587 25.2.4 Mercury found in Fluorescent Light Bulbs is not the Predominant Source of Mercury that may Enter the Environment as a Result of Light Bulb Use and Disposal 588 CONTENTS xvii 25.2.5 What if We would have Examined the Full Life Cycle Impacts of MTBE Before it was Commercialized to Reduce Smog in Cities? 590 25.2.6 Quality and Safety are Imperative Considerations in the Design and Development of Every Product Made Today, but It was not Always so 591 25.2.7 Geographical Information Systems (GIS) were Initially Expensive and Data Collection was Time Consuming, but Today GIS Systems are Commonplace in Most Planning and Decision Support Functions 592 25.2.8 In the 1970s, Carnival Led the Way in Making Cruising Affordable for the Masses 592 25.3 Diagnostics and Interpretation 593 25.4 Concluding Remarks 595 References 596 597 Index Preface For a growing number of companies, global diversity is a business imperative. Manufacturing operations have increasingly become technically and geographically diverse in the sourcing of resources, manufacturing and assembly operations, usage, and final disposal. Thisexpansion, along with a growing awareness of sustainability and the responsibilities to the environmental, economic, and social dimensions that go with it, has prompted environmental managers and decision makers everywhere to look holistically, from cradle to grave, at products and services. The need for a tool that helps users obtain data and information to accurately and consistently measure the resource consumption and environmental aspects of their activities has never been more acute. Most importantly, people now realize that decisions should not lead to improving one part of the industrial system at the expense of another. In other words, the identification and avoidance of unintended consequences are essential in the decision making process. Out of this need came Life Cycle Assessment (LCA). What started as an approach to compare the environmental goodness (greenness) of products has developed into a standardized method for providing a sound scientific basis for product stewardship in industry and government. When used within an environmental sustainability framework, LCA ultimately helps to advance the sustainability of products and processes as well as promote society's economic and social activities. When I set out to create the "latest and greatest" book on Life Cycle Assessment (LCA), I had three very specific goals in mind. First, I wanted it to be comprehensive, covering every possible facet of methodology and application. This was quite a challenge, given the ever-growing scope that LCA has reached over the years. As can be seen in the table of contents, the subject is addressed from a wide range of perspectives and in many applications. Note, however, that this book is not a "how to" manual with step-by-step instructions for conducting an LCA. Instead, I designed this book to explain what LCA is, and, just as importantly, what it is not. The immense popularity of the "life cycle" concept led to its use in a variety of assessment approaches, even in those approaches that are focused on a single environmental aspect. For example, LCA is often used in writing about carbon accounting. In these times of heightened concern over climate change, indi- viduals and organizations alike are eager to measure the release and impact of greenhouse gases. But the results only address climate change and not the other equally important impacts. The exact meaning of the methodology is frequently misunderstood, resulting in carbon footprint and LCA being used synonymously, and incorrectly so. By narrowing an assessment to a single issue of concern, the results will not reflect the important benefit that LCA offers of identifying potential xix xx PREFACE trade-offs. There are several other similar examples, which I will not go into here. I trust that after reading this book, the differences will be clearer. Second, I wanted the reader to hear from the experts and leaders in LCA. I asked recognized LCA professionals for their contributions. I felt it was important to hear all the representative voices from industry, academia, and of course, the LCA con- sultants. We even heard from non-governmental organizations (NGOs). The book contains writings from 47 authors from 10 countries. Despite their busy schedules, all of the authors came through with marvelous contributions. I give my sincere thanks to the authors for their dedication and hard work and their willingness to take time away from their extremely busy careers and lives to share their experiences, wisdom, observations, and guidance which made this book possible (the term "herding cats" was used frequently as I waited for final manuscripts). In the end, I am extremely pleased with the outcome. There is much the reader can learn by drawing from the wealth of experience and knowledge that is contained within the covers of this book. Third, I wanted to capture the latest advancements in LCA methodology and application in one convenient place. I also wanted to indicate where further advancement in LCA is still needed. The book was designed with a particular flow in mind. It begins at the beginning, with an historical account of LCA and how it has developed over the years. The following chapters cover the basics of the LCA methodology, and discuss goal and scope definition, inventory analysis, impact assessment, and interpretation. Then, multiple examples of application are presented. This is followed by aspects of how LCA is used in decision making, and how it is now evolving as the underlying principle behind environmental sustain- ability. The book is best approached from beginning to end, as each chapter was designed to build on the last. However, each chapter is self-contained, and readers may benefit from skipping to the topic(s) of interest to them. LCA and LCA-based tools give us a way to improve our understanding of the environmental impacts associated with product and process systems in order to support decision making and achieve sustainability goals. In the early 1990s (before the first ISO 14000 series on LCA was established), there was considerable confusion regarding how LCA should be conducted. Even the term itself was debated, and 'life cycle analysis' and 'life cycle assessment' were used interchangeably. Eventually, 'assessment' became the preferred choice in the ISO standards and within the LCA community. 'Analysis' is still used by some (usually those who are less familiar with LCA), but I asked the authors to use 'assessment' throughout their writing to be consistent with the ISO standard, and to appease me. Over the last 22 years, it has been fascinating to watch the evolution of LCA practice, from concept to standard- ized methodology and on to being the 'backbone' of sustainability. I intend for this book to be a useful reference tool for a wide audience, including students in environmental studies, government policy makers, product designers and manufacturers, and environmental management professionals. That is, I hope it is useful to anyone who wants to implement a life cycle approach in their orga- nization, be it in the private sector or public, as well as those who simply wish to have a better understanding of what all the fuss over LCA has been about. Cincinnati, Ohio, USA July 2012 Mary Ann Curran 1 Environmental Life Cycle Assessment: Background and Perspective Gjalt Huppes1 and Mary Ann Curran2* institute of Environmental Sciences (CML), Leiden University, Leiden, The Netherlands 2US Environmental Protection Agency, Cincinnati, OH, USA Abstract Life Cycle Assessment (LCA) has developed into a major tool for sustainability decision support. Its relevance is yet to be judged in terms of the quality of the support it pro- vides: does it give the information as required, or could it do a better job? This depends very much on the questions to be answered. The starting point was the application to rel- atively simple choices, such as making technical changes in products and choosing one material over another, with packaging as a main example. This was then followed by the use of LCA in consumer choices. Over time, there has been a shift to more encompassing questions, such as the attractiveness of biofuels and the relevance of lifestyle changes. This chapter describes the ongoing discussions on issues that still need to be addressed, such as allocation, substitution data selection, time horizon, attributional versus conse- quential, rebound mechanisms, and so forth. The chapter then describes how LCA might develop in the future. There are important tasks ahead for the LCA community. Keywords: Life cycle assessment, LCA, allocation, attributional, consequential, decision support 1.1 Historical Roots of Life Cycle Assessment The concept of exploring the life cycle of a product or function initially developed in the United States in the Fifties and Sixties within the realm of public purchasing. Back then, usecost often carried the main share of the total cost. A first mention of the life cycle concept, by that name, is by Novick (1959) in a report by the RAND Corporation, focusing on Life Cycle Analysts of cost. Costs of weapon systems, a main application at that time, include not only the purchasing cost, or only the use cost. They also cover the cost of * The views expressed in this chapter are those of the authors and do not necessarily reflect the views or policies of the US Environmental Protection Agency. Mary Ann Curran (ed.) Life Cycle Assessment Handbook: A Guide for Environmentally Sustainable Products, (1-14) © 2012 Scrivener Publishing LLC 1 2 LIFE CYCLE ASSESSMENT HANDBOOK development and the cost of end-of-life operations. Life Cycle Analysis (not yet referred to as 'Assessment') became the tool for improved budget man- agement, linking functionality to total cost of ownership. This was a first for government. Method issues and standardization questions soon followed. How should data on past performance be related to expected future perfor- mance? How is functionality defined? Can smaller systems like jet engines be taken out of overall airplane functioning? Should system boundaries encompass activities such as transport? How should accidents and mistakes be considered? How should overhead costs and multi-function processes be allocated? For public budget analysis, the life cycle approach led to gen- eral questions on methodology and standardization, as in Marks & Massey (1971), also linking to other "life cycle-like' tools for analysis, especially cost-benefit analysis. The life cycle concept rapidly spread to the private sector where firms struggled with similar questions. By 1985, a survey paper (Gupta & Chow, 1985) showed over six hundred explicit life cycle studies that had been pub- lished, all focusing on relating system cost to functionality. The methodol- ogy issues were treated in an operational manner, for example by Dhillon (1989). Optimizing system development and system performance became a core goal for the now broadly applied public and private life cycle analysis of cost. There is now over a half a century of experience with function-based life cycle analysis of system costs, see the survey in Huppes et ah (2004), continu- ing in parallel with environmental Life Cycle Assessment, or environmental LCA (moving now from 'Analysis' to 'Assessment'), and later to the life cycle concept related to Life Cycle Costing (LCC). Returning to these roots might be an interesting endeavor. 1.2 Environmental Life Cycle Concepts This life cycle concept was already fully developed when environmental policy became a major issue in all industrialized societies, at the end of the Sixties and in the early Seventies. Environmental policies, mainly command- and-control type, were at first source-oriented with very substantial reduc- tions in emissions being realized. It soon became clear that such end-of-pipe measures were increasingly expensive. However, other options were not eas- ily introduced into the mainly command-and-control type regulatory frame- work as it had been developed. Shifts in mode of transport, for example, were clearly of broad environmental importance, but not easily brought into the regulations. The comparative analysis of such different techniques for a simi- lar function was hardly developed in a practical way. Cost-Benefit Analysis (CBA), as an example, was focused at projects that aim to maximize welfare. It was made obligatory for environmental regulatory programs in the US, starting in 1971 with Executive Order 20503, on Quality of Life. Adapted substantially by consecutive US presidents, it still is a main contender for ENVIRONMENTAL LIFE CYCLE ASSESSMENT 3 environmental LCA in the public domain applications, and increasingly so in the European Union (EU) as well. Environmental LCA first developed rela- tively unobserved by the private sector, before having the name shortened to simply "LCA" at the end of the Eighties. Both CBA and LCA have a life cycle concept at their core. The major difference between them is that CBA speci- fies activities in time and then uses a discounting method, in line with domi- nant modes of economic analysis, which is similar to the Life Cycle Analysis of cost. LCA, on the other hand, uses a timeless steady-state type of sys- tem analysis, without discounting effects. CBA also quantifies environmen- tal effects in economic terms and then discounts them. In modeling welfare effects of climate policies, for example, the discounting mode is dominant. That dynamic analysis seems superior to the static GWP (Global Warming Potential) analysis used in LCA. How to quantify environmental effects in an economic sense and how to discount effects spread across time remains a core issue in CBA, open to further public and scientific debate. In LCA the time frame discussion is hardly present. Looped processes are not, and can- not, be specified in time. The only explicit treatment of time is found in the consideration of the different environmental themes in GWP impacts, with scores being limited to 20, 50 or 100 years, and in the toxic effects of heavy metals and the like that are assumed to extend virtually to eternity. The time frame discussion, then, might be part of Interpretation, which is problematic in itself while also hardly any guidance is given in the ISO standards or in any of the instructional guides that followed. It would be interesting to have a discourse on overlapping issues and stra- tegic choices in the domains of Cost-Benefit Analysis; Life Cycle Analysis of costs; and environmental Life Cycle Assessment. 1.3 LCA Links to Environmental Policy The conceptual jump from life cycle cost analysis to the first life cycle-based waste and energy analysis, and then to the broader environmental LCA (how we view LCA today) was made through a series of small steps. Documented history starts with the famous Coca Cola study from 1969, see Hunt and Franklin (1996), who were involved in LCA right from that start. The environ- mental focus was on resource use and waste management, not yet the broad environmental aspects that are usual in LCA now. The broad conceptual jump to environmental LCA as contrasted with Life Cycle Analysis of cost was made in the Eighties and formalized in the Nineties with the work of SETAC and the standardization in the 14040 Series of ISO, see Klöpffer (2006). From the start with the RAND Corporation in the end of the Fifties, the system to be analyzed was clear. It should cover the supply chain, including research and develop- ment, the use stage, and the processing of wastes from all stages, including end-of-life of the product analyzed. The link to public policy was made based on concepts first developed in the Netherlands, in the Eighties at the Department of Environmental 4 LIFE CYCLE ASSESSMENT HANDBOOK Management headed by Pieter Winsemius. After the first stage of environ- mental policy, with command-and-control instruments directed at main sources, there was a shift to a systems view, and to a more general formula- tion of environmental policy goals in the Dutch Environmental Policy Plans, see also Winsemius (1990, original 1986). This shift from a source-oriented to an effect-oriented approach created a domain for environmental LCA from an environmental policy point of view, as contrasted to a business long-term cost view or a consumer interest point of view. Winsemius coined the envi- ronmental themes approach now dominant in LCA, looking for integration over the environmental compartments policies regarding water, air and soil. His overall policy strategy was based on now familiar themes: Acidification; eutrophication; diffusion of (toxic) substances; disposal of waste; and dis- turbance (including noise, odour, and local-only air pollution). Somewhat later, further national policy themes were added:climate change; dehydra- tion; and squandering. The theme-oriented policy formed the basis for a broadened view on envi- ronmental policy, now covering complementary entries like volume policy, product policy and substance policy. In their implementation it was no lon- ger only chimneys and sewers but also people and organisations: the target groups of environmental policy, several groups of producers and consumers. The responsibility for consequences of actions shifted to these target groups, which had to internalise the goals of environmental policy as specified using the themes approach. If, how, and why this internalization happened is a sub- ject of much debate; see de Roo (2003) for a first analysis. For doing so, the new metrics of the themes were most appropriate, indicating the environmental performance of business and consumers in a unified collective framework, that of (generalized) public environmental policy. Private organizations may have ideas on what themes should constitute the impact assessment. It is the col- lective point of view that creates the relevance of LCA outcomes. The themes approach remained specifically Dutch for a short while only. It inspired envi- ronmental policy of the EU; see the historic survey by Liefferink (1997). It was incorporated in LCA in an operational manner beginning in the Nineties, as the Life Cycle Impact Assessment method now dominant in LCA, of course with additions and adaptations. In the US the themes approach was not dominant in environmental policy, with more emphasis there on CBA. That probably was the reason that the introduction of the themes approach in environmental LCA followed later there. It is an open question now if and how Life Cycle Impact Assessment can be linked to environmental themes as goals of public policy. These goals might be - but need not be - the goals of a specific country or of the EU. Public policy goals set as targets, for example as emission reduction targets for a substance, lack the integrative power of the themes approach. Goals set as general wel- fare maximation lack the link to specific domains of action. Themes can make the link. Also because product systems and LCA increasingly become global, passing the policy goals of specific countries, the foundations for the themes in LCA impact assessment should be clarified. ENVIRONMENTAL LIFE CYCLE ASSESSMENT 5 1.4 Micro Applications of LCA Rising The last decades have seen a startling rise in the production of LCAs. There are consultants in virtually all countries, many with an international orientation. Databases and software have become widely available. There also are interest- ing in-firm developments. Two Netherlands-based firms we happen to know have their internal LCA capacity well developed, Philips and Unilever. Procter and Gamble contributes a chapter to this book on their LCA operations. The Unilever example is enlightening. They regularly produce internal LCAs on virtually all of their products, having produced well over a thousand LCAs by now. They use the LCAs for product system improvement, reducing easily avoidable impacts. These may seem tiny per product, but may be substantial from a dynamic improvement point of view. Tea bags used to have zinc plated iron staples to connect the bag and the carton handle to the connecting thread. This gave a dominant contribution to the overall life cycle impact of the tea bag system. The staples were first replaced by a glue connection and in many cases now by a sewing connection. Such product system improvement forms the core of LCA use. However, when having so many equivalent LCAs, new more strategic applications become possible. Can strategies be developed to reduce environmental impact covering more than one product, with more general guidelines for product development? Such applications are now developing in Unilever, see the box. Similarly, Philips has developed strategic guidelines at an operational level regarding the use of materials, reducing the number in each product and phasing out those with the largest contribution to environ- mental impacts. LCA, in its micro level application, is now a two decade-old success story. With all caveats following, we should not throw out the baby with the bath water. LCA is here to stay, and the child is still growing. 1.5 The Micro-Macro Divide The core goal of environmental LCA as was established in the Nineties was to help improve environmental quality, with product policy - internalized, private, and also in public regulations - as one entry into environmental pol- icy. That role is based on the assumption that improved micro environmental performance of a product-function system corresponds to an environmental improvement at the macro level. That macro level in principle is global society at large in its environmental impacts, as product systems increasingly span the world. When looking at the mechanisms that link shifts or developments in micro level behavior to macro level performance it is perfectly clear that there is no direct correspondence. Cycling as mode of transport has a minor fraction of the impacts of car transport per kilometer traveled, but also has a minor fraction of the costs. Some elements of this discrepancy may be covered by eco-efficiency analysis of these transport systems, expressing environmental impacts not per functional unit but per Euro spent. Such micro level scores 6 LIFE CYCLE ASSESSMENT HANDBOOK don't tell what the ultimate outcome of a shift to cycling in commuting will be. The income not spent on cars will be spent on something else, anything. The shift to cycling is also linked to a different spatial infrastructure, with different retail systems, different housing requirements, etc. Though one may be con- fident that this is all to the environmental good - there may be good reasons to believe so - that assessment is not just based on LCA. The analysis of the overall system effect can easily be set up in a way that cycling really is bad. If the income not spent on cars is assumed to be spent to a substantial degree on flight based holidays, the net environmental outcome of more cycling might well be negative. When reckoning with such behavioral mechanisms, the choice of mechanism will determine the outcome, quite haphazardly at the moment. So the question is if a strategy for analysis can be set up to include the most relevant mechanisms in an equitable way. The move towards conse- quential LCA is a possible step, but not the only one. A core question is if dynamic, non-linear mechanisms can be incorporated in the comparative static or steady state framework of LCA, as consequential LCA. Or, should the micro level LCA technology system better be placed in a broader modeling system reckoning with income effects, dynamic market mechanisms, structural effects and constraints, and what more might be rel- evant? The modeling required definitely does not fit in the linear homogenous system of LCA based on matrix inversion for easy solutions. It seems wise to first investigate divergent cases with an open mind as to most relevant causali- ties, and to look into options for structured modeling later. Then a choice for micro-type consequential LCA might be substantiated, or not, or only for some applications. 1.6 Macro Level LCA for Policy Support The use of LCA in public policy has been coming up, with an LCA-type of analysis being used. The domain of application of LCA has been that of spe- cific product choices. However, the link to broader policy issues, never absent, seems on the rise. Biofuel, see below, is a major example, with unresolved dis- cussion in the EU. The general feature of policy applications is that they should show how a change considered would work out, requiring an ex ante analy- sis of consequences of policy options, or an ex post analysis showing how a policy has worked out. In both cases we need toknow 'how the world would have been different/ The functional unit with an arbitrary volume then is to be replaced by an analysis covering the total volumes. Policies tend to be set up in order to reach specified goals, not marginal effects of an unknown volume. Using traditional arbitrary-unit LCA for policy support assumes a correspon- dence between micro level LCA outcomes and macro level consequences for the choice at hand. This assumption should be substantiated. It also relates to the average versus marginal discussion, with causalities most easily estab- lished at a marginal level, but overall effects then requiring integration over all marginal changes, as increments. For substantiating the consequences of ENVIRONMENTAL LIFE CYCLE ASSESSMENT 7 the policy choice at hand, the technical relations as covered in LCA should be part of the analysis, but also the broader behavioral mechanisms should be covered. If all mechanisms together come out negative, showing a rebound, simple LCA would have given the wrong advice. A first step for the analysis is to place the choice in a framework of totals for society. Input-output analysis with environmental extensions can be set up in an LCA-type manner, with some details added to better cover the choice at hand. This hybrid analysis has come up as a theoretical tool, with one applica- tion related to the option of using fuel cell buses in urban transport, see Cantono et ah (2008). In the old Life Cycle Analysis of cost, the same link to input-output analysis was pointed out previously, see Staubus (1971). This IO framework allows one to specify one first secondary effect, the income effect. The higher cost of fuel cell buses replacing Diesel buses implies lower spending on other items, with lower environmental impacts there. However, this IO-analysis is static and cannot cover well broader causal mechanisms. Causal analysis can only be specified in time. It is the before-after analysis, of the situations with-and- without specific alternative policies. So the second step involves a dynamic anal- ysis, of all mechanisms leading to the overall, the macro level, consequences. The conclusion is that for supporting policy choices with macro level con- sequences the arbitrary functional unit based LCA will often be too narrow to give valid answers. A broader framework for analysis is then required. 1.7 Example Biofuels In the biofuels discussion, all levels of questions come up. They range from small-step improvement options for a given biotechnology to produce biofu- els; to the comparison between different fuels, including biofuels; and to an evaluation of a global shift towards a more biobased energy system. When looking at a small system, one may assume the changes to be so small that indirect effects are negligible. But the sum of all these small changes adds up to a substantial change. A small change in biomass demand for energy will have a small effect on biomass production and a small effect on energy prices. However, such effects are additive, and often non-linearly increasing. If biofuel is relevant, it has to be produced in substantial amounts. This also holds for the minor improvement in biotechnology. So, indirect effects cannot be ignored. A next option for simplified analysis is the assumption that all mechanisms not covered remain equal or do not influence the outcome. Both assumptions generally are not true in the case of bio-energy, see the OECD (Organization for Economic Cooperation and Development) study by Doornbosch and Steenblik (2008). These should be investigated empirically. A final option is to make assumptions on the rest of the world. One may assume, for example, that all additional biomass will come from barren lands not fit for food producing agriculture. This assumption is often present in studies on second and third generation biofuels. However, the use of fertile grounds will mostly be cheaper than barren grounds to produce biomass - that is why they were barren. In 8 LIFE CYCLE ASSESSMENT HANDBOOK general, no mechanism exists to restrict biomass production for fuel to barren lands only. Therefore, to develop sound advice on biofuel choices we have to be comprehensive and cover 'all relevant mechanisms/ What might these relevant mechanisms be for biofuels? A first set of mecha- nisms relates to the markets more or less directly involved. In the US case of corn based ethanol (first generation) or stover-based ethanol (second genera- tion), this involves the fodder and food markets for these products. Directly connected are other products for these markets, especially wheat. Also directly linked are changes in land use, more corn and wheat pressing out other sta- ple products like soy beans, increasing the price of soy beans a well. These three staple crops function on global markets, so even if the bioethanol is US-produced the effects are really global, in principle affecting all crops glob- ally. The overall agricultural effect will include somewhat higher prices, an intensification of agriculture, with also higher nitrous oxides emission affect- ing climate, and an increase in the volume of agricultural land use. Two studies have investigated the impact on additionally induced conversion of tropical rainforest into agricultural land; see Searchinger et ah (2008) and Fargione et ah (2008). These two studies differ in set-up and outcomes and cannot directly be connected to LCA-type studies. They show however that such global effects of biofuel production cannot be neglected. One mechanism not covered by these studies is a feedback in spatial policy as has taken place in Brazil and Indonesia, with strengthened legislation and strengthened power in imple- mentation. This administrative reaction to US, and similar EU, biofuel policy will of course have longer term effects mainly. Some of these issues will be treated in a bit more detail by Guinee in a later chapter, as the framework for Life Cycle based Sustainability Analysis (LCSA). So here we are, with old-fashioned types of LCA studies showing how attractive biofuels may be, and a range of induced mechanisms often being detrimental in an environmental sense, both on the shorter, longer and very long term. What to do? The only answer seems to be: get on the job, make a framework for analysis, start filling in the framework with conceptual models, and produce first order quantifications on environmental outcomes. On the way to specifying the mechanisms involved one will encounter major social effects as well, with rising food prices in cities (with riots and possibly a major effect on the uprisings in the Middle East) and with rising agricultural incomes all over the world, also for the poorest farmers. How to come to an overall evaluation of several divergent effects spread out in time will be a next prob- lem to solve, involving all problems that have already been encountered in Cost Benefit Analysis, but often have not been not solved adequately yet. 1.8 Why Environmental LCA? The early development of cost-oriented LCA had clear goals: reducing cost while improving performance. That driver remains, with cost analysis an essential element in management accounting. ENVIRONMENTAL LIFE CYCLE ASSESSMENT 9 Decision making on product systems developed in a period when planning and control was the dominant mode of organisation, with the "owner" having control over the supply chain. However, in the period that LCA emerged the planning concepts started to shift. When discussing chain management with the environmental officer of Nokia, there was a startling reaction: "How can we?" Nokia had a policy to have all their suppliers renewed through com- petitive bids every three months. But supply chain developments did not stop there. With globalisation large firms now tend to outsource production, with also chain management outsourced. Big brands do development and market- ingand shift the production to the chain as much as possible, to reduce costs and risks. Exceptions are where consumers are the driving force for LCA, as with branded consumer products, especially food, and with special NGO action-based cases, like Nike. The link with environmental-based public policy then becomes weaker, also because such theme- based approaches are less frequent in public policy making now, or at least shifting to new themes. There is a tendency to shift the analysis to applied subjects dominant in public discussion, like resources, energy, waste and land use. These however are not environmental impacts, but aspects of concern, for several reasons, including environmental ones. Resources are a supply problem mostly related to costs and to market failure, leading to nationalist policies to safeguard supply. The underlying depletion aspect, leading to increasing environmental impacts in primary production, would have been covered by the themes approach as in terms of acidification, climate change, etc. Similarly, energy is a concern also because of costs and supply security, hardly being environmental issues. There are more fossil fuel resources available than the earth can accommodate pleasantly for mankind. The climate issue was there already, and it still is. It does not require a special energy impact, although, of course, energy use plays a dominant role in the cli- mate problem. Depending on the way exergy and heat are produced and used, and on the volumes involved, there will be environmental impacts in terms of the LCA themes, including climate change. The specification of 'waste' now tends to include waste to-be-processed, to focus on recycling issues, implying a system boundary not with the environment. When looking at the environmen- tal issues covered in the public discussions on firms, there is a clear tendency to shift to domain-specific indicators in the chain, again leaving the principle of system definition to cover all processes to linking to the environment. The Global Reporting Initiative (GRI, 2012) covers the environmental reporting of firms, especially multinational ones. They do not adhere to a themes approach and tend to apply indicators 'within the system/ to use tra- ditional LCA terms. For example, there are customized indices for sectors like financial services, electric utilities, NGOs, food processing, mining & metals, airport operators, and construction & real estate. The reporting is to cover what to external parties is deemed relevant. There is no well-defined con- ceptual basis for specifying such concerns. The once globally dominant posi- tion of the EU in conceptualizing environmental policy seems to have been eroded. The environmental themes approach is no longer the dominant mode 10 LIFE CYCLE ASSESSMENT HANDBOOK of goal setting in environmental policy in the EU, and it never has been in the US and Japan. This shift in public and private concerns may have severe impacts on the development of LCA. With the impact assessment shifting to in-system subjects, system boundaries become less clearly defined, and the environ- mental issues of the themes approach are not covered by 'full system analy- sis/ The overall analytics of the impact assessment in terms of midpoints categories and endpoint categories is left with the new impacts of energy, resources and waste, very similar to the old ones of early LCA in the US in the Seventies. Box 1.1 The Role of Life Cycle Assessment in Unilever (a Personal Account) Back in 1991, Chris Dutilh, then Development and Environmental Manager of Unilever's margarine company in the Netherlands (then called Van den Bergh & Jürgens) got approval from his Managing Director to hire some- one to do Life Cycle Assessment (LCA). The undersigned was the lucky man. LCA in those days was an emerging concept. SETAC (Society for Environmental Technology and Chemistry) had taken the lead in method- ology development. Application of LCA in food and agriculture had not really been done yet. But in the Netherlands the first Covenant on Packaging had been signed between industry and government. The Covenant called for voluntary reductions (on the part of industry) of packaging volumes used, and LCA seemed an appropriate tool to analyse various options. Between 1991 and 1993,1 had the pleasure of conducting various LCAs of different packaging systems. Unilever Netherlands had taken the lead on the product group mayonnaise, mustard, jams and dressings (anything in jars and bottles that was not a drink). We used LCAs to investigate various options to reduce weight of glass jars (always good), switch from glass to plastics (not always straightforward), or change the cap material (very dependent on recycling options of the cap material). Glass recovery and recycling was already at rather high level in the Netherlands at the time (over 80% of packaging glass recovered, if memory serves me well). We also investigated a re-use scenario. One way glass packaging had been introduced some fifteen odd years before. The result of our scenario study (which was supported by the Unilever engineering department, to get the details of our glass jar washing plant as realistic as possible) was that reuse of glass jars had considerable environmental benefit, but also serious cost implications, mostly labour cost. In other words: negative environmental impact could be reduced by putting more labour in the supply chain. The work on LCA moved to Unilever R&D in Port Sunlight, UK, and became a department within Unilever's Safety and Environmental ENVIRONMENTAL LIFE CYCLE ASSESSMENT 11 Assurance Centre (SEAC). Over the years, dozens of LCA's of Unilever products and supply chains were done. This experience came in handy when Unilever prepared its Sustainable Living Plan (launched in November 2010, see http://www.sustainable-living.unilever.com/). In the plan, Unilever focuses on carbon footprints and water footprints across the entire value chain. In order to calculate a baseline (2008), we performed LCAs on 1600 representative products, combined with consumer use data of 14 countries around the world. The combined outcome represents about 70% of Unilever's sales value (which was 45 billion Euro in 2010). It showed us that of our carbon footprint, only 3% is in our own factories: 26% is upstream, with suppliers of our raw materials, 2% is in transport throughout the value chain, 1% is in post-consumer waste disposal, and a whopping 68% is in consumer use in the household. Performing LCAs has become part of everyday decision making in Unilever. We have learned many lessons, as a result. Dr. Jan Kees Vis Global Director Sustainable Sourcing Development Unilever R&D Vlaardingen 1.9 Overview of the Book As we already mentioned, the last several decades have seen a dramatic rise in the application of LCA in decision making. The interest in the life cycle concept as an environmental management and sustainability tool continues to grow. This book was created to concisely and clearly present the various aspects of LCA in order to help the reader to better understand the subject. The content of the book was designed with a certain flow in mind. After a high level overview to describe current views and state-of- the-practice of LCA, it presents chapters that address specific LCA method- ological issues. These are followed by example applications of LCA. Finally, the book concludes with chapters that link LCA and responsible decision making and how the life cycle concept is a critical element in environmental sustainability. 1.9.1 Methodology and Current State of LCA Practice The book continues with an "Overview of the Life Cycle Assessment Method - Past and Future" in which Heijungs and Guinee describe at a conceptual level the methodology and current state of LCA practice. The chapter also explores present developments that are influencing the evolvingmethod. Detailed discussions on methodology are given in the chapters by Sauer on life cycle inventory (LCI) and by Margni and Curran on life cycle impact assessment (LCIA). 12 LIFE CYCLE ASSESSMENT HANDBOOK Life Cycle Assessment (LCA) relies heavily on both data and software. Reliable data is the driving force behind LCA as large amounts of process and production data are needed. The chapter by Curran on sourcing inven- tory data discusses historical and current practices in sourcing LCI data and proposes futuristic approaches for reporting process inventory data, including manufacturer self-reporting, using open-source models. Ciroth explores cur- rently available LCA software and highlights the current status and trends for LCA software into the future. 1.9.2 LCA Applications Through a range of case studies, authors explore how typical methodological issues have been treated and managed in various example applications. Of growing interest is how to model bio-based systems. In "Modeling the Agri- Food Industry with LCA" Notarnicola, Tassielli, and Renzulli emphasize the need for a harmonized framework for conducting food-related LCAs and for collecting and reporting data for agri-food chains in both agricultural and industrial applications. Landers, Urban and Bakshi note many engineering analyses undervalue or completely ignore the ecosystem goods and services that are essential to all human activities, such as fresh water, soil, carbon and nitrogen cycles, and pollination, and propose a framework that more accurately accounts for them. They present a case study that compares different ecosystem services using exergy and emergy analysis and highlight the importance of "Accounting for Ecosystem Goods and Services in LCA and Process Design." In exploring how Fortune 100 companies can better manage the supply chain and improve a product manufacturer's sustainability metrics, Weisbrod and Loftus of Procter and Gamble present "A Case Study of the Practice of Sustainable Supply Chain Management." P&G's sustainable supply chain management program, through collaboration with supply chain partners, enabled the company to link environmental sustainability and social responsi- bility with business operations and values. Two chapters look closely at specific aspects of materials management throughout the life cycle. Weitz discusses "End of Life Materials Management" and how taking a life-cycle perspective encourages waste planners to consider the environmental aspects of the entire system including activities that occur outside of the traditional activities of waste disposal. Similarly, but at the other end of the life cycle, environmental impacts of the mining and minerals pro- cessing sectors are often inadequately reflected in LCAs. In "Application of LCA in Mining and Minerals Processing" Stewart, Holt and Rouwette describe how LCA is being used in the mining and minerals sector and indicate where LCA needs to be refined to meet the needs of the industry. Other areas of LCA application are provided in chapters on forest prod- ucts by Bolin; building systems by Todd; product innovation by daSilva; food waste and packaging by Hanssen, Moller, Svanes and Schakenda; and sustainable chemistry & engineering by Hunter, Helling, and Shiang. ENVIRONMENTAL LIFE CYCLE ASSESSMENT 13 1.9.3 LCA Supports Decision Making and Sustainability Subsequent chapters then broaden the scope of the book by exploring how LCA can be integrated with economic and social aspects of sustainability to provide a deeper analysis that encompasses relevant dynamic mechanisms. In this vein, Potting, Gheewala, Bonnet, and van Buuren look at four assess- ment methods associated with human health and environmental impacts (Technology Assessment, Environmental Impact Assessment, Risk Assessment, and LCA) to provide guidance to stakeholders on when to use what assess- ment method. Prado, Rogers, and Seager also give a critical eye to the interpre- tation of impacts, specifically how normalization and valuation are applied in the decision making process. Benoit Norris dedicates a chapter to the newer, fast growing area of social LCA methodology, and why it should be of interest to decision makers along with environmental assessment approaches. Building on social LCA and con- sidering the cost aspect along with LCA, Zamagni, Guinee, Heijungs and Masoni present a framework for "Life Cycle Sustainability Analysis/' LCSA is intended to deepen the scope of analyses by integrating physical, social, economic, cultural, institutional and political considerations into the decision making process. Stevenson and Ingwersen explore environmental product claims that range from simple product characteristic claims made by manufac- turers to those based on full LCA with additional metrics 1.9.4 Operationalizing LCA The final chapters offer a look at the role that life cycle information, in the hands of companies, governments and consumers, may have in improving the environmental performance of products and technologies LCA practitioners in developed countries struggle to keep up with demand of their services. Developing countries and emerging economies are even less capa- ble of harnessing the potential in LCA for sustainable development. In "Building Capacity for Life Cycle Assessment in Developing Countries" Toolsearam addresses the critical issue of building a critical mass of mass of people with the right capacities in LCA in less developed regions of the world. Internationally, the success of the sustainability paradigm needs the par- ticipation of many stakeholders, including citizens, corporations, academia, and NGOs. Governments in particular play a very important role with the leverage they have through procurement, regulation, international treaties, tax incentives, public outreach, and other policy tools. In "Environmental Accountability: A New Paradigm for World Trade is Emerging," Ngo pres- ents her view of a shifting world paradigm where LCA is the foundation of decision-making in regulation and commerce, and poses a number of oppor- tunities and challenges. And finally, Fava provides personal reflections on how "Life Cycle Information Informs Greener Products." He points to a trend for incorpo- rating life cycle information into the design and development processes for 14 LIFE CYCLE ASSESSMENT HANDBOOK products and policies, just as quality and safety concerns are now addressed throughout product design and development. He cautions that while recent trends suggest that integration of LCA into all manner of decision making will continue to increase, we must act, by providing education and improved tools and databases, to ensure that it does. References Cantono, S., R. Heijungs and R. Kleijn (2008). "Environmental accounting of eco-innovations through environmental input-output analysis: the case of hydrogen and fuel cell buses/' Economic Systems Research 20(3): 303-318. Dhillon, B.S. (1989). Life Cycle Costing: Techniques, Models and Applications. New York: Gordon and Breach. Doornbosch, R., and R. Steenblik (2007). Biofuels: is the cure worse than the disease? Round Table on Sustainable Development, OECD, Paris. Available at: http://www.oecd.org/datao- ecd/9/3/39411732.pdf, accessed January 2012. Executive Order 20503 (1971). Quality of Life. Executive Office of the President Office of Management and Budget, Washington 20503, October 5,1971. Accessed 13 January 2012 at: http: / / www.thecre.com/ombpapers/QualityofLifel .htm. Fargione, J., J. Hill, D. Tilman, S. Polasky, and P. Hawthorne (2008). "Land Clearing and the Biofuel Carbon Debt." Sciencexpress, 7 February 2008, science. 1152747. Fiorello, Marco R. (1975). Problems in Avionics Life-Cycle Analysis. Santa Monica, CA: RAND Corporation, 1973. http://www.rand.org/pubs/papers/P5136. Also available in print form. Gupta, Yash, and Wing Sing Chow (1985). "Twenty-FiveYears of Life Cycle Costing - Theory and Applications: A Survey." International journal of Quality & Reliability Management, Vol. 2, Issue 3, pp.51-76. Hunt, Robert G., William E. Franklin, and R.G. Hunt (1996). "LCA- How it came about. Personal reflections on the origin and the development of LCA in the USA." International Journal of Life Cycle Assessment Vol. 1, Nr 1, pp.4-7, DOI: 10.1007/BF02978624. Huppes, Gjalt, Martijn van Rooijen, Rene Kleijn, Reinout Heijungs, Arjan de Koning, and Lauran van Oers (2004). Life Cycle Costing and the Environment. Available at: http://www.rivm.nl/ milieuportaal/images/Report%20LCC%20April%20%202004%20final.pdf (accessed January 2012). Klöpffer, Walter (2006). "The Role of SETAC in the Development of LCA." International Journal of Life Cycle Assessment, Volume 11, Supplement 1,116-122, DOI: 10.1065/lca2006.04.019. Liefferink, Duncan (1997). "The Netherlands: a net exporter of environmental policy concepts." Chapter 5 in: Mikael Skou Andersen and Duncan Liefferink (1997). European environmental policy; the pioneers. Manchester: Manchester University Press. Marks, K.E. and H.G. Massey (1971). Life Cycle Analysis Procedures and Techniques: an Appraisal and Suggestions ForFuture Research. Santa Monica, CA: Rand Corporation, ADA132027. Novick, David (1959). The federal budget as an indicator of government intentions and the implications of intentions. Santa Monica, CA: Rand Corporation, publication P-1803. A summary is in the Journal of the American Statistical Association, Vol. 55, No. 290, June 1960. Roo, Gert de (2003). Environmental planning in the Netherlands: too good to be true. From command- and-control planning to shared governance. Farnham: Ashgate, ISBN: 978-0-7546-3845-2. Searchinger, T., R. Heimlich, R.A. Houghton, F. Dong, A. Elobeid, J. Fabiosa, S. Tokgoz, D. Hayes, and T.H. Yu (2008). "Use of U.S. Croplands for Biofuels Increases Greenhouse Gases Through Emissions from Land Use Change." Sciencexpress, 7 February 2008, science.1151861. Staubus, G. (1971). Activity Costing and Input-Output Accounting. Homewood, IL: Richard D. Irwin. Winsemius, Pieter (1990). Guests in our own home: thoughts on environmental management. McKinsey & Company. Translation of: Gast in Eigen Huis (1986). Alphen aan den Rijn: Samsom H.D. Tjeenk Willink. 2 An Overview of the Life Cycle Assessment Method - Past, Present, and Future Reinout Heijungs and Jeroen B. Guinee Institute of Environmental Sciences (CML), Leiden University, Leiden, The Netherlands Abstract This chapter gives an overview of the mainstream method behind Life Cycle Assessment (LCA). It does so on the basis of the generally accepted principles, canon- ized by the International Organization for Standardization (ISO). The first part of the chapter is an overview devoted to the method itself and the current state of the prac- tice. The second part provides a sketch of the historical development that led toward the method. The chapter concludes with a description of present developments that are influencing the evolving method. Keywords: Life cycle assessment, method, ISO, life cycle inventory, life cycle impact assessment 2.1 The Present-Day LCA Method Life Cycle Assessment (LCA) refers to the process of compiling and evaluat- ing the inputs, outputs and the potential environmental impacts of a product system throughout its life cycle [1]. But LCA also refers to the result of this pro- cess. In this chapter, we will focus on the process, i.e. the method that is used to obtain an "LCA result/7 LCA has come a long way, and it continues to change. Since a decade or so ago, there has been a broadly accepted set of principles that can be claimed as the present-day LCA framework. This section reviews this LCA framework, without going into depth, and without going into the newer developments. The International Organization for Standardization (ISO) has issued a series of standards and technical reports for LCA, referred to as the 14040 series. This series consists of the documents listed in Table 2.1. Mary Ann Curran (ed.) Life Cycle Assessment Handbook: A Guide for Environmentally Sustainable Products, (15-42) © 2012 Scrivener Publishing LLC 15 16 LIFE CYCLE ASSESSMENT HANDBOOK Table 2.1 ISO documents on life cycle assessment (LCA). Number 14040 14041 14042 14043 14044 14047 14048 14049 Type International standard International standard International standard International standard International standard Technical report Technical report Technical report Title Principles and framework Goal and scope definition and inventory analysis Life cycle impact assessment Life cycle interpretations Requirements and guidelines Examples of application of ISO 14042 Data documentation format Examples of application of ISO 14041 Year 1996,2006 19981 20001 20001 20062 2003 2001 2000 1 Updated in 2006 and merged into 14044. 2 Replaces 14041,14042, and 14043. Life cycle assessment framework Goal and scope definition Inventory analysis Impact assessment Interpretation Direct applications: - Product development and improvement - Strategic planning - Public policy making - Marketing - Other Figure 2.1 The general methodological framework for LCA (ISO 14040 [1]). AN OVERVIEW OF THE LIFE CYCLE ASSESSMENT METHOD 17 The standards are organized into the different phases of an LCA study. These are: • Goal and scope definition • Inventory analysis • Life cycle impact assessment • Life cycle interpretation The relationships between these phases have been illustrated in a figure, and this figure has become a type of logo of LCA (Figure 2.1). Typically, LCA starts by defining goal and scope, then proceeds to the inven- tory analysis, then optionally continues to impact assessment, and it ends with the interpretation. However, as indicated in Figure 2.1, an LCA study is a highly iterative process, so that the LCA practitioner may need to go back to goal and scope after the preliminary inventory work, to move back from impact assessment to inventory analysis, to have a look at the interpretation in an early stage, etc. Below, we will discuss the main idea and content of the four phases in sep- arate subsections. All quotations are taken from the ISO documents, unless otherwise indicated. 2.1.1 Goal and Scope Definition There is no explicit ISO definition of the first phase of LCA. However, it obvi- ously centers around formulating the question and stating the context of answering this question. In the goal and scope definition, no data is collected and no results are calculated. Rather, it is a place where the plan of the LCA study is defined as clearly and unambiguously as possible. Likewise, in an LCA report, it should help the reader to quickly find out the precise question addressed and main principles chosen. The goal of the LCA should deal with the following topics: • the intended application; • the reasons for carrying out the study; • the intended audience; • whether the results are to be used in comparative assertions disclosed to the public. The choices made here have an influence on the rest of the LCA procedure. For instance, depending on the intended audience, a critical review may be needed, and it may be important that an external expert takes this task. In the scope definition, a number of major choices are made. First of all, the product system or systems to be studied, and the function the system delivers (or in case of a comparative LCA, the functions the systems deliver). For instance, one might be interested in the product systems incandescent light bulb versus the LED bulb, with the function of lighting a room. 18 LIFE CYCLE ASSESSMENT HANDBOOK An important aspect of the scope definition is the functional unit. It is obvi- ously pointless to compare an incandescent light bulb with anLED light bulb: the life spans and performances differ considerably, and the function is not having a light bulb, but having light of a certain quality. The functional unit expresses the function of the products, and, thereby, offers a way to equalize differences in performance. A functional unit for analyzing lighting systems could thus better be phrased in terms of the function, for instance "lighting a standard room of 15 square meters with 1000 lumen for 1 hour." As LCA math- ematically employs a linear calculation rule, the results will scale by choosing a numerically different functional unit (say, "lighting a standard room of 20 square meters with 800 lumen for 3 hours"), but the alternatives consid- ered will scale up or down consistently, so this will not affect the conclusions. A consequence is, however, that LCA cannot tell if a product is "sustainable" or "environmentally friendly;" LCA can only indicate if product X is "more sustainable" or "more environmentally friendly" than product Y, or that the use phase is the "least sustainable" or "least environmentally friendly" part of the life cycle for product Z. The scope definition further sets the main outline on a number of subjects that are discussed and further refined in more detail in the later phases. These include, among others: • system boundaries; • impact categories; • treatment of uncertainty. The ISO standard and some other text in places suggest that these topics are implemented in detail in the scope definition. This is wrong: the goal and scope definition is not concerned with collecting data or calculating results, so no concrete details on such topics can be specified at this phase. 2.1.2 Inventory Analysis ISO defines life cycle inventory analysis (LCI) as the "phase of life cycle assess- ment involving the compilation and quantification of inputs and outputs for a product throughout its life cycle." It will be clear that quantification is an important aspect here, and numbers, in terms of data and calculations, are of central concern in the inventory analysis. The LCI is built on the basis of the unit process. A unit process is the "smallest element considered in the life cycle inventory analysis for which input and output data are quantified." Examples of unit process are coal mining, steel production, refining of oil, production of furniture, use of a television set, recy- cling of waste paper, and transport by lorry. Each of these processes is described in quantitative terms as having inputs and outputs. In LCA, a unit process is treated as a black box that converts a bundle of inputs into a bundle of outputs. Inputs come in several types: products (including components, materials, and services), waste for treatment, and natural resources (including fossils, ores, A N OVERVIEW OF THE LIFE CYCLE ASSESSMENT METHOD 19 From other unit processes From the environment To other unit processes To the environment Figure 2.2 General template of a unit process. The process (grey rectangle) is considered as a black box, having inputs (left-hand side) and outputs (right-hand side) from and to other unit process (top arrows) and from and to the environment (bottom arrows). Electricity production Semi- L conductor production TV production TV use Figure 2.3 Fragment of a simplified flow diagram for an LCA on television (TV) sets. Because the purpose is to show how unit process are connected, only the flows from and to other unit processes are displayed, and flows from and to the environment are hidden. All transport, packaging, etc. has been left out as well. biotic resources, and land). Outputs come in several types as well: again prod- ucts (including components, materials, and services), waste for treatment, and residuals to the environment (including pollutants to air, water, and soil, waste heat, and noise); see Figure 2.2. Unit processes form the building blocks of an LCA. This is because products are not harmful for the environment as such, except for the processes involved in products. Producing, using, and disposing products creates the burden to the environment. Therefore, these processes assume a central position in LCA. The essential feature of LCA in which it distinguishes itself from the analysis of an industrial or agricultural process is that it connects different unit process into a system. A flow diagram is a graphical representations of the system com- prised of connected unit processes. Figure 2.3 show a fragment of such a flow diagram. As we can see, some unit processes are connected with one another in simple upstream-downstream connections, e.g., TV production is upstream connected to semi-conductor production. But there are also more complicated connections, e.g., electricity linking to different parts of the system, and recy- cling feeding back to production. Flow diagrams are in fact huge webs of inter- connected unit processes. In the present era of digital databases, LCA studies can easily comprise several thousands of unit processes. 20 LIFE CYCLE ASSESSMENT HANDBOOK LCA is primarily a quantitative model. In the LCI, all unit processes included have to be quantified. This means that we have to specify the sizes of the inflows and outflows, per unit process. As an example, let us take the unit process of aluminum production. An aluminum plant may specify their technology in term of inputs and out- puts by stating input requirements (e.g., 2 kg aluminum oxide and 20 kWh electricity per kg produced aluminum) and emissions (e.g., 2 g dust per kg produced aluminum). We must translate this into our template for unit pro- cesses; see Table 2.2. For each of the unit processes included, quantitative data should be collected. Moreover, in order to be able to process the data and perform the calculations automatically, a clear and unambiguous representation is needed. This implies, among other things, harmonization of nomencla- ture (e.g., not using "carbon dioxide" for one unit process and "C0 2 " for another), and harmonization of units (e.g., not mixing up kilograms (kg) and pounds (lb)). In Table 2.2, the unit process data is given per unit of output, here per kg of aluminum. In an LCA, we must next find out how much we need. For instance, the product may need 3 kg of aluminum, not 1 kg. The basic assumption of the LCA model is that technologies are linear. This means that we can scale the data of a unit process by a simple multiplication. In the example, 3 kg of aluminum would require 6 kg of aluminum oxide and 60 kWh of electricity, while it would release 6 g of dust. The assumption of linear technology is an important restriction of LCA; yet it is an important step in making the calcula- tion and data collection feasible. In scaling the unit processes, the web-like nature of the system quickly cre- ates complications, as everything depends upon everything. The calculation of the scaling factors, and with that of the emissions to and extraction from the environment, is greatly simplified by considering the problem as a system of linear equations: one unknown (the scaling factor) for every unit process, and one equation (a balance) for every flow. Thus, solutions may be obtained by matrix algebra. The details of this are not discussed here; see [2] for a detailed exposition. The approach mentioned above may fail in a number of cases. We mention two complications: Table 2.2 Example of unit process specification in an aluminum plant. Type of Flow inputs from other unit processes inputs from other unit processes outputs to other unit processes outputs to the environment Name aluminum oxide electricity aluminum dust Amount 2 20 1 0.002 Unit kg kWh kg kg AN OVERVIEW OF THE LIFE CYCLE ASSESSMENT METHOD 21 • For some products, upstream production processes or down- stream disposal processes may be difficult to quantify; • For someunit processes, the balance equations become impossible due to the fact that these processes produce not just one product but several co-products. The first issue can be solved by a procedure known as cut-off, the second one by allocation. Cut-off is a solution to the problem that the system is theoretically infinitely large. To produce a TV, we need machines, and these machines are produced by machines, and these machines in turn need machines, etc. But of course we have an intuitive idea that some very distant upstream processes will be quite unimportant to the study. This means that we will cut-off certain inputs: although we know that something is needed, and we sometimes even know how much is needed, we do not go into the trou- ble of specifying how these inputs are produced. It turns out to be diffi- cult to specify reliable criteria when cut-off is allowed, or to estimate how large the error is when a cut-off is made. Criteria on the basis of negligi- ble contribution to mass or cost (e.g., smaller than 1%) often work pretty well, but occasionally have been shown to yield large errors. Alternatively, estimates of missing parts by means of similar processes (e.g., estimating the production of a freezer by using production data for a refrigerator), or by economic input-output tables may be helpful. Another approach is to conduct a difference analysis: in comparing a cathode ray tube (CRT) with liquid crystal display (LCD) television we may leave out the broadcasting processes. The second problem, co-product allocation, has given rise to one of the biggest controversies in LCA theory. The example problem can be demon- strated thus: If a transportation process needs gasoline, the upstream unit pro- cess is a refinery that produced not only gasoline, but also diesel, kerosene, heavy oils, and many other products. The direct impacts (from pollutants like C02), but also the flows to and from other processes that may lead to impacts (e.g., from oil drilling) may be argued not be attributable to gasoline only, but need to be distributed over gasoline, diesel, and all the other co-products. This is hardly contested, but the debate focuses now on how to do this. To make the issue more concrete, the question at hand can be stated as: How much of the C02 from a refinery is allocated to the gasoline? Different schools have provided different arguments, none of which have been completely compel- ling so far. Some solutions lead to strange results, while other solutions may be very difficult to carry out (e.g., for lack of data or appropriate software). Still others are rejected outright by many experts. To complicate the issue, the prob- lem does not only occur in unit process that produce several co-products, but also in unit processes that treat more than one type of waste, or where waste is recycled into a useable good. It is not even agreed upon if the multi-output case, the multi-input case, and the recycling case must be treated in the same way or not. 22 LIFE CYCLE ASSESSMENT HANDBOOK Within ISO, a preferred order for solving the multi-functionality problem has been designed. It distinguishes several solutions (dividing the unit process into two or more sub-processes, expanding the system to include the addi- tional functions, partitioning on the basis of a physical parameter, partitioning on the basis of an economic parameter), separated by clauses like "wherever possible" and "where ... cannot be established." This stepwise procedure is a clear comprise, and in practice leaves so much freedom that LCA studies that are according to the ISO standard can give conflicting results. One peculiarity deserves to be mentioned: besides the ISO-based "expanding the system to include the additional functions," we often see a method that is best described as "subtracting the avoided impacts from additional functions," but that is more commonly known as the substitution method or the avoided burdens method. For instance, when a waste treatment activity co-produces electricity, the emissions from the regular way of producing the same amount of electric- ity are subtracted. This method has similarities with that of system expansion, but of course they are not identical. Many LCA studies employing the substi- tution method claim to be ISO-compliant, even though strictly speaking ISO 14044 does not mention this method, let alone recommend it. That does not necessarily mean that these studies are incorrect, of course. Compliance with ISO is not a sufficient quality guarantee, but also not a necessary one. After appropriate cut-off and allocation steps, the final inventory results can be calculated. Typically, this is a table with the quantified inputs from and outputs to the environment, for each of the alternative systems considered, expressed in relation to the functional unit. With the present-day software and databases, this inventory table may be 1000 lines or longer. It contains not only the familiar pollutants and resources, such as C02, ΝΟχ, and crude oil, but also more exotic items, such as 1-pentanol, cyprodinil, and dolomite. 2.1.3 Impact Assessment Life cycle impact assessment (LCIA), or impact assessment for short, is the "phase of life cycle assessment aimed at understanding and evaluating the magnitude and significance of the potential environmental impacts for a prod- uct system throughout the life cycle of the product." Its motivation comes from two observations: • The final result of the inventory analysis, the inventory table, is too long (e.g., 1000 different items) to handle; • The inventory table contains many items that require expert knowledge (such as 2-methyl-2-butene) to understand in terms of importance. Impact assessment, and in particular the characterization step, solves both issues: it "involves the conversion of LCI results to common units and the aggregation of the converted results within the same impact category." A N OVERVIEW OF THE LIFE CYCLE ASSESSMENT METHOD 23 While the unit process is the central element of the inventory analysis, the central element in impact assessment is the impact category. ISO defines it as a "class representing environmental issues of concern to which life cycle inven- tory analysis results may be assigned." Perhaps more helpful are some exam- ples: climate change, toxicity, and depletion of fossil energy carriers. As climate change (often used interchangeable with global warming) is a well-known issue, we will illustrate the main ideas of impact assessment with this case. The inventory table contains a number of greenhouse gases: C02, CH4, N 2 0 , etc. These are known to contribute all to the phenomenon of cli- mate change. Climate change involves long sequence of causal mechanisms: emissions of greenhouse gases lead to changes in the composition of the atmo- sphere, which lead to a change in the radiation balance, which in turn leads to a change in the temperature distribution, which leads to changes in climate, which leads to changes in ecosystems and human activities, etc. The further we proceed in this causal chain, the more uncertain and speculative our knowledge becomes. While quite some scientific evidence is available with respect to the composition of the atmosphere, the impacts on biodiversity are debated. Many of these later impacts are even conditional on our future activities, including future emission scenarios and mitigating actions. To be able to quantitatively model the emissions of different greenhouse gases into an impact indicator for climate change, we must do several things. First, we must choose a certain point in the causal mechanism. This can be at the front end (change in radiation balance), at the back-end (change of bio- diversity), or somewhere in between (change in temperature). In LCA, two main schools have emerged: • Those that focus on the front-end, the so-called midpoint approach; • Those that focuson the back-end, the so-called endpoint approach. The midpoint approach has the advantage that it includes fewer debatable assumptions and less-established facts; the endpoint approach has the advan- tage that it provides more intuitive metrics (like loss of life years instead of kg C02-equivalents). Regardless of the choice between midpoint and endpoint, the indicator chosen is referred to as the impact category indicator, or category indicator for short. Second, a way must be found to convert the emission data into the cho- sen impact indicator. Scientists in chemistry, meteorology, ecology, etc, have developed model fragments to estimate the atmospheric life-times of greenhouse gases, their effect on the radiation balance and the forma- tion of clouds, the effects of temperature on the distribution of species, etc. These fragments have been combined by workgroups from the UN-based International Panel on Climate Change (IPCC) into quantitative models of the impacts of greenhouse gas emissions. Part of this is the global warm- ing potentials (GWPs), which are quantitative measures of the strength of different greenhouse gases. Many midpoint LCI A methods apply GWPs 24 LIFE CYCLE ASSESSMENT HANDBOOK for climate change. We will illustrate their usage below. For now, it suffices to say that GWPs provide one example of a set of characterization factors, and that the IPCC-model from which they are derived is an example of a characterization model. Note, by the way, that IPCC has not developed this model as a characterization model for LCIA, but that the LCA-community has adopted this model as such and its derived GWPs as characterization factors. Also note that the characterization model itself is not used by LCA practitioners; only the characterization factors that have been derived from it as a one-time exercise are used. Characterization factors are often tabu- lated in LCA guidebooks and are implemented in many LCA software pack- ages, while the characterization models often require supercomputers and expert knowledge. In fact, one element is needed before one can select a category indicator and a characterization model with associated characterization factors. It is the selection of impact categories to be addressed. Some LCA studies concentrate on just one impact category. For instance, the carbon footprint (of a product, not of a company or country) is considered a form of LCA that addresses just climate change at the midpoint level through GWPs. At the other extreme, some LCA studies incorporate fifteen or more impact categories. For con- sistency reasons, the choice of impact categories is often made on the basis of a recommended impact assessment guidebook or its implementation in software. Thus, in practice one often sees LCA-studies reporting the use of "IMPACT2002+," "TRACI," "CML-IA," "ReCiPe," "ILCD," etc. All these methods comprise a recommended set of impact categories with a category indicator and set of characterization factors. ISO does not specify any choice in these matters. Table 2.3 gives an overview of some often-used impact catego- ries and category indicators. We see that the column with endpoint indicators contains many times the same term (e.g., "loss of life years"). This suggests that impact categories can be aggregated into fewer endpoint indicators than midpoint indicators. As a concrete example of how characterization works, let us study a fragment of a hypothetical inventory table, containing the following information: emis- sion of C02100 kg, emission of CH41 kg, emission of S0 2 1 kg. Characterizing greenhouse gases with GWPs requires a table with GWPs. In such a table, one can find that the GWP of C0 2 is 1 (by definition) and that the GWP of CH4 is 25 (kg C02-equivalent/kg CH4). S02 has no GWP; it is assumed not to contribute to climate change. Characterization now proceeds in the case of climate change by calculating 1 x 100 + 25 x 10 = 350 kg C 0 2 - equivalent For the more general case, this can be written as GW = ^GWPsxms A N OVERVIEW OF THE LIFE CYCLE ASSESSMENT METHOD 25 Table 2.3 Overview of widely-used impact categories with examples of category indicators at Midpoint and Endpoint levels. Impact Category climate change ozone layer depletion acidification eutrophication human toxicity (sometimes split into carcinogenics, non-carcinogenics, respiratory effects, etc.) eco-toxicity (sometimes split into aquatic toxicity, terrestrrial toxicity, marine toxicity, etc.) depletion of energy carriers depletion of material resources land use impacts water use impacts Midpoint Category Indicator infra-red radiative forcing change in tropospheric ozone concentration H+ concentration biomasss potential time-integrated exposure, corrected for hazard time-integrated exposure, corrected for hazard primary energy requirement amount of material used, corrected for availability and/or importance amount of land occupied or transformed amount of water used or displaced Endpoint Category Indicator loss of life years, fraction of disappeared species loss of life years fraction of disappeared species fraction of disappeared species loss of life years fraction of disappeared species decreased availability decreased availability fraction of disappeared species decreased availability where GW is the global warming score, s the substance (the different green- house gases), GWPs the GWP of substance s, and ms the emitted amount of substance s in kg. This may be further generalized as s where c stands for the impact category, I represents the indicator result for category c, and CFcs is the characterization that links substance s to impact 26 LIFE CYCLE ASSESSMENT HANDBOOK category c. This formula is the operational formula for characterization. With a table of characterization factors specified, it makes clear that: • LCIA builds on the results of LCI (as is clear from the term ras); • Characterization converts the results of LCI into a common metric (as is clear from the multiplication by CF); • Characterization aggregates the converted LCI results (as is clear from the summation symbol). The result from characterization is a list of numbers, for instance a score for climate change, a score for toxicity, etc. ISO refers to such numbers as "category indicator results/' but most LCA practitioners prefer names like "score/7 sometimes expanded with the name of the impact (as in "toxicity score"). The complete list is known by names such as "LCIA profile," "characterization table," etc. After this sketch of the principle of characterization, let us have a look at the more formal ISO point of view. Impact assessment in ISO is structured into a number of steps: • Selection of impact categories, category indicators and character- ization models; • Classification; • Characterization; • Normalization; • Grouping; • Weighting; • Data quality analysis. Although characterization is just one of the steps of impact assessment, the former term is often used as a pars-pro-toto for the latter term. Indeed, in discussing the principle of characterization above, we have touched upon the steps of selecting categories, indicators, and models and characterization. These are mandatory steps for ISO. The step of classification is mandatory as well, but few LCA studies report it. ISO defines it as the "assignment of LCI results to the selected impact categories." Its purpose is to clearly show which emissions and extractions are treated under which impact category, but it involves no numerical conversion into a common metric, as is the case for characterization. Normalization refers to calculating "the magnitude of the category indi- cator results relative to some reference information." It is an optional step for ISO, and indeed, many LCIA studies stop at the characterization. The reference informationis in most cases that total impact in a certain region in a certain time period, e.g., in the country of decision in one year. Normalization helps "to understand better the relative magnitude for each indicator result." Without normalization, the indicator results are in quite different units, e.g., kg C02-equivalent for climate change and MJ primary energy for fossil energy AN OVERVIEW OF THE LIFE CYCLE ASSESSMENT METHOD 27 depletion. To put these results in perspective, the normalization expresses them as a share of the total impact size in the region. Arbitrary differences due to a choice of units disappear, and it becomes clear to which impact category a product contributes relatively much. The units of the normalize indicator results are equal; nevertheless, such numbers cannot meaning- fully be added because the severity of the different impact categories has not yet been accounted for. This can be done in the weighting step; see below. Normalization fulfills several functions: it provides insight into the meaning of the impact indicator results, it helps to check for errors, and it prepares for a possible weighting step. Grouping is another optional step, although it is seldom seen in LCA studies. ISO defines it as "the assignment of impact categories into one or more sets." ISO mentions two ways: • Sorting (on a nominal basis, like global/regional/local); • Ranking (on an ordinal basis, like high/medium/low priority). Weighting, like characterization, converts and aggregates, but while charac- terization does so for the LCI results, weighting starts with the characteriza- tion (or normalization) results. Typically, weighting factors are applied, either to the characterization indicator results, or to their normalized version. The weighting factors themselves are supposed to reflect value judgements, such as social and political priorities. Weighting typically produces one final num- ber, by means of: W = £ W F c x 7 c c where Ic again symbolizes the impact score (or normalized impact score) for impact category c, WFc the weighting factor for this impact category, and W the weighted result. Well-known examples of such weighted results are the eco-indicator and the ELU (environmental load unit). The data quality analysis, finally, relates to an analysis of completeness, uncertainties, etc. We treat this more systematically in the section on interpre- tation below. 2.1.4 Interpretation ISO defines interpretation as the "phase of life cycle assessment in which the findings of either the inventory analysis or the impact assessment, or both, are evaluated in relation to the defined goal and scope in order to reach conclu- sions and recommendations." Several elements are mentioned by ISO: • Identification of significant issues; • An evaluation that considers completeness, sensitivity and consistency checks; 28 LIFE CYCLE ASSESSMENT HANDBOOK • Conclusions, limitations, and recommendations. • Appropriateness of the definitions of the system functions, the functional unit and system boundary; • Limitations identified by the data quality assessment and the sensitivity analysis. The text of ISO on interpretation is very concise, and no details are given on procedures and techniques to be employed. The same applies to most guide- books on LCA. They mention carrying out an uncertainty analysis, but give no clear guidance on how this should be done. In another context, we have introduced the distinction between procedural and numerical approaches [3]: • Procedural approaches include all types of analyses that deal with the data and results in relation to other sources of information, like expert judgement, reports on similar products, intuition, reputation of data suppliers, and so on. • Numerical approaches include those approaches that somehow deal with the data that is used during the calculations, without reference to those other sources of information, but as algorithms that use and process the data in different ways, so as to produce different types of "smart" data reduction that provide an indication of reliability, key issues, discernibility, robustness, and so on. This distinction helps us understand some important roles of interpretation. On the one hand, it is about comparing the data and results with previous findings, and putting the results in the context of decision-making and limita- tions. On the other hand, it is devoted to a systematic analysis with the help of statistical and other decision-analytic techniques. The latter type may be incorporated in software, and indeed, an increasing number of software pack- ages contain options for running Monte Carlo analysis, doing sensitivity anal- ysis, carrying out statistical significance tests, etc. For instance, in the CMLCA software, we have implemented, among others: • Contribution analysis; • Comparative analysis; • Uncertainty analysis; • Perturbation analysis; • Key issue analysis; • Discernibility analysis. The iterative nature of the ISO framework (Figure 2.1) shows up in this con- text. Whenever the uncertainties are too high, we may go back to collect better data. Whenever sensitivity analysis shows that some decisions are crucial, we may go back and do a more refined analysis. In this way, the interpretation helps to prepare for a balanced decision, while helping improve the LCA. AN OVERVIEW OF THE LIFE CYCLE ASSESSMENT METHOD 29 The practice of LCA is quite meagre, unfortunately. We still see many LCA studies without uncertainty or sensitivity analysis, even though software increasingly facilitates this. There is of course a psychological argument that a contractor pays for finding out something, not for increasing the doubt. And as many LCA practitioners spend several months on collecting data, it is never a nice thing to waste this effort in a last-minute uncertainty analysis. But decision-making obviously means also taking into account the limits of knowl- edge. Moreover, as discussed before, a proper analysis of uncertainties and sensitivities helps to prioritize the steps earlier on in the framework: collecting data, setting boundaries, making choices. 2.1.5 LCA in Practice In the text above, the emphasis has been on the generally accepted practice. This is a mix of the ISO standards and a not precisely defined set of guidebooks. These latter include the ILCD Handbook, the Dutch Handbook on LCA, the guidelines from EPA, those from JEMAI, and others. All these texts inter- pret, add, refine, or modify the ISO standards. As has been indicated at a few places, the practice in LCA is sometimes different from what the ISO stan- dards prescribe. There are differences in terminology (e.g., one seldom sees the term intermediate products), in method (cf. the frequent use of the substitu- tion method), in quality control (judged by frequent absence of uncertainty analyses), etc. There are also de facto additional standards, dictated by the use of soft- ware and databases. Many software packages for LCA have built-in options for impact assessment and uncertainty analysis, but nearly always in a restricted form, allowing some variants and prohibiting other variants. LCI databases are often constructed with pre-defined allocation methods and cut-off rules, so the user cannot choose otherwise, and cannot carry out sen- sitivity analyses. One last, important aspect is the incorporation of new insights. The ISO standards date from the period 1997-2000, and the 2006 update is not really an update (just a merging). Meanwhile, numerous developments have taken place. We mention just a few, without discussing their meaning or importance: • The distinction between attributional and consequential LCA; • The development of input-output-based LCA and hybrid LCA; • The incorporation of economic and behavioural mechanisms in the LCA model; • The development of new impact categories, e.g., for land use andbiodiversity; • The development of life cycle costing and social LCA, and their fusion into life cycle sustainability analysis (LCSA); • The application of LCA to things other than products, like policies and life styles. 30 LIFE CYCLE ASSESSMENT HANDBOOK Some of these developments are already becoming mainstream LCA by now, while other developments may disappear. Anyhow, the practice of LCA is evolving, and it will continue to do so. The next part of this chapter will discuss some of these developments in more detail. 2.2 A Short History of LCA Above, we saw that there is not one LCA, despite the standards set by ISO. The standardization process itself had to deal with a field that had already grown into a highly diverse patchwork. The ISO-standards served well in merging many ideas in a common framework and providing a terminology. The con- sensus-based ISO process could propose a framework and terminology, but it could not provide detailed guidelines, let alone data on unit processes and characterization factors. The freedom that the ISO standards offer is in some respects a curse, as it leads to different "ISO-compliant" reports on the same topic with contradict- ing results. But on the other hand, the fact that there is freedom has increased its acceptance by practitioners and researchers. Meanwhile, these researchers do not accept a fossilized LCA, but are developing and maturing LCA further. This section discusses the historical development of LCA in terms of its past, present, and future. The text is largely based on and adapted from [4] (Guinee et al., Life Cycle Assessment: Past, Present, and Future, Environmental Science & Technology 45,1,90-96. Copyright 2011 American Chemical Society). 2·2.1 Past LCA (1970-2000): Conception and Standardization In this section we will briefly discuss and evaluate LCA as developed and applied in the past, while distinguishing two periods: (1) 1970-1990 and (2)1990-2000. 2.2.2.2 1970-1990: Decades of Conception The first studies that are now recognized as (partial) LCAs date from the late 1960s and early 1970s, a period in which environmental issues like resource and energy efficiency, pollution control and solid waste became issues of broad public concern [5]. The scope of energy analyses [6,7,8], which had been con- ducted for several years, was later broadened to encompass resource require- ments, emission loadings and generated waste. One of the first (unfortunately unpublished) studies quantifying the resource requirements, emission loadings and waste flows of different beverage containers was conducted by Midwest Research Institute (MRI) for the Coca Cola Company in 1969. A follow-up of this study conducted by the same institute for the U.S. Environmental Protection Agency in 1974 [9] and a similar study conducted by Basier & Hofman [10] in Switzerland, marked the beginning of the development of LCA as we know it today. The MRI used the term Resource and Environmental Profile Analysis AN OVERVIEW OF THE LIFE CYCLE ASSESSMENT METHOD 31 (REPA) for this kind of study, which was based on a systems analysis of the production chain of the investigated products "from cradle to grave." After a period of diminishing public interest in LCA and a number of unpublished studies, there has been rapidly growing interest in the subject from the early 1980s on. In 1984 the Swiss Federal Laboratories for Materials Testing and Research (EMPA) published a report [11] that presented a comprehensive list of the data needed for LCA studies, thus catalyzing a broader application of LCA [5]. The study also introduced a first impact assessment method, dividing airborne and waterborne emissions by semi-political standards for those emis- sions and aggregating them, respectively, into so-called "critical volumes" of air and "critical volumes" of water. The period 1970-1990 comprised the decades of conception of LCA with widely diverging approaches, terminologies and results. There was a clear lack of international scientific discussion and exchange platforms for LCA. During the 1970s and the 1980s LCAs were performed using different methods and without a common theoretical framework. LCA was repeatedly applied by firms to substantiate market claims. The obtained results differed greatly, even when the objects of the study were the same, which prevented LCA from becoming a more generally accepted and applied analytical tool [12]. 2212 1990-2000: Decade of Standardization The 1990s saw a remarkable growth of scientific and coordination activities world-wide, which is reflected in the number of workshops and other forums that have been organized in this decade [13,14,15,16,17,18] and in the number LCA guides and handbooks produced [19,20,21,22,23,24,25,26]. Also the first scientific journal papers started to appear in the Journal of Cleaner Production, in Resources, Conservation and Recycling, in the International Journal of LCA, in Environmental Science & Technology, in the Journal of Industrial Ecology, and in other journals. Through its North American and European branches, the Society of Environmental Toxicology and Chemistry (SETAC) started playing a lead- ing and coordinating role in bringing LCA practitioners, users and scientists together to collaborate on the continuous improvement and harmonization of LCA framework, terminology and methodology. The SETAC "Code of Practice" [27] was one of the key results of this coordination process. Next to SETAC, the International Organization for Standardization (ISO) has been involved in LCA since 1994. Whereas SETAC working groups focused at development and harmonization of methods, ISO adopted the formal task of standardization of methods and procedures. The period of 1990-2000 can therefore be characterized as a period of convergence through SETAC's coordination and ISO's standardization activi- ties, providing a standardized framework and terminology, and platform for debate and harmonization of LCA methods. In other words, the 1990s was a decade of standardization. Note, however, that ISO never aimed to standard- ize LCA methods in detail: "there is no single method for conducting LCA." 32 LIFE CYCLE ASSESSMENT HANDBOOK During this period, LCA also became part of policy documents and legisla- tion. The main focus was on packaging legislation, for example, in the EU [28] and the 1995 Packaging Law in Japan [29]. Although LCA has proven its value in these policy-based applications, there were also problems with respect to the authoritativeness of results (cf. [30,31]). Several well-known life cycle impact assessment methods, still used today, evolved from methods developed in this period, such as the CML 1992 environmental theme approach [22,26], endpoint or damage approaches [32,33] but also the nowadays broadly accepted [34,35] multi-media approach for assessing potentially human and ecotoxic emissions [36]. Although this decade is mainly one of convergence, it is also the stage of scientific scrutiny, research into the foundations of LCA, and exploring the connections with existing disciplines. For instance, we observe sprouting ideas on consequen- tial LCA and related allocation methods [37,38,39]. These and other sophisti- cations mark the transition to the present decade of LCA, which is a decade of elaboration but also of divergence in methods again. 2.2.2 Present LCA (2000-2010): Decade of Elaboration The first decade of the 21st century has shown an ever increasing attention to LCA. In 2002, the United Nations Environment Programme (UNEP) and the Society for Environmental Toxicology and Chemistry (SETAC) launched an International Life Cycle Partnership, known as the Life Cycle Initiative [40]. The Life Cycle Initiative's main aim was formulated to put life cycle thinking into practice and improve the supporting tools through better data and indicators. Life cycle thinking also continuedto grow in importance in European Policy, as highlighted through, e.g., the Communication from the European Commission of the European Communities (CEC) on Integrated Product Policy [IPP; 41]. On top of this, life cycle thinking was also incorpo- rated in, e.g., the thematic strategies on the Sustainable Use of Resources [42] and on the Prevention and Recycling of Waste [43]. In its 2003 Communication on Integrated Product Policy (IPP), the European Commission underlined the importance of life cycle assessment and the need for promoting the application of life cycle thinking among the stakeholders of IPP [41]. In response, the European Platform on Life Cycle Assessment [44] was established in 2005, mandated to promote the availability, exchange, and use of quality-assured life cycle data, methods and studies for reliable decision support in (EU) public policy and in business. In the USA, the U.S. Environmental Protection Agency started promoting the use of LCA [45]. Various national LCA net- works were also established like, for example, the large-scale Australian LCA Network [46] and the American Center for LCA [47], both in 2001, and the smaller scale Thai network [48] in 2000. In this same period, environmental policy gets increasingly life-cycle based all over the world (e.g., [49,50]). For example, several life cycle-based carbon footprint standards have been, or are being, established [51]. This A N OVERVIEW OF THE LIFE CYCLE ASSESSMENT METHOD 33 standardization for environmental policy raised serious problems, which have not yet been solved adequately [52]: • As life-cycle based carbon footprint calculations may constitute the basis for decisions, e.g. granting subsidies to stimulate the use of bio-energy, it is of utmost importance that the indicator results be robust and lawsuit-proof/ This implies that the freedom of methodological choices for the handling of data, e.g., biogenic carbon balances and allocation, should be reduced to an absolute minimum [53]. • Another topic is that the limited scope of carbon footprints is not sufficiently accounted for when using the results. The scopes of carbon footprint studies can be limited in geographic coverage (dominated by Europe and North America), in feed stocks covered, in the number of different emissions to the environ- ment included, and in environmental impacts addressed (carbon footprint studies are typically limited to global warming, while other environmental impacts can be more important when assess- ing the sustainability of products, for example, biofuels: eutrophi- cation, acidification, ecotoxicity and human toxicity, biodiversity, water use, etc.; [54]). These limitations should at least be clearly reported as part of the conclusions of current, narrow-scope carbon footprint studies. • A final topic of concern is the translation from functional-unit-based to real-world improvements. This may be the most difficult issue to address. Side-effects such as indirect land use, rebound effects, market mechanisms, and such all play a role in how large-scale production of biofuels would affect the food market, scarcity, social structure, land use, nature and other things that are impor- tant for society. These are insufficiently addressed by current LCA studies, as was identified and analyzed by Sheehan [55], van der Voet and Lifset (li), and in the EU FP6 CALCAS project [56]. Although consequential LCA (e.g., [57]) is very strong in map- ping impacts of indirectly affected processes of a decision, mod- eling macroscopic land use changes on the basis of microscopic consequential product LCAs (bottom-up) is not likely to result in long-run sustainability. It may be more realistic to start think- ing how more realistic, macroscopic scenarios for land use, water, resources and materials, and energy (top-down) such as drafted by the IPCC [58] and in the work by Graedel and van der Voet [59] can be transposed to microscopic LCA scenarios. The period 2000-2010 can be characterized as the decade of elaboration. While the demand on LCA increases, the current period is characterized by a divergence in methods again. As ISO never aimed to standardize LCA methods 34 LIFE CYCLE ASSESSMENT HANDBOOK in detail and as there is no common agreement on how to interpret some of the ISO requirements, diverging approaches have been developed with respect to system boundaries and allocation methods [60,61], dynamic LCA [62,63,64,65], spatially differentiated LCA [60,61], risk-based LCA [66,67,68,69], environmental input-output based LCA (EIO-LCA) based and hybrid LCA [70,71,72] that may have a tense relation with some of the basic principles of the ISO standards. On top of this, life cycle costing (LCC; cf. [73]) - first used in the 1960s by the US Department of Defense for the acquisition of high- cost military equipment [74] - and social life cycle assessment (SLCA; cf. [75]) approaches have been proposed and/or developed that may have consistency problems with environmental LCA in terms of system boundaries, time per- spectives, calculation procedures etc. [76,77]. These different approaches have the life-cycle basis in common but they differ in the methodological elaboration and in the question(s) they are address- ing. We need to clarify exactly how the various approaches differ or overlap, but, most importantly, we need to clarify the link between questions and approaches: which approach is useful for which question. Despite new LCA textbooks being published [78,79,80], there is a further need for structuring this varying field of LCA approaches. We also need to take into account more types of externalities (economic and social impacts) and more mechanisms (rebound, behavior, price effects, dynamics) to meet the above-mentioned shortcomings of existing LCA studies in the field of, for example, biofuels while meeting specific user needs such as in simplified LCA. The European Commission acknowledged this challenge and commissioned the CALCAS (Co-ordination Action for innovation in Life Cycle Analysis for Sustainability) project in 2006 to structure the varying field of LCA approaches and to define research lines and programmes to further LCA where necessary. The CALCAS project has been finished and results have been published [81]. One of its main results concerns the establishment of a framework for Life Cycle Sustainability Analysis (LCSA) linking life cycle sustainability questions to knowledge needed for addressing them, identifying available knowledge and related models, knowledge gaps and defining research programmes to fill these gaps. 2.2.3 Future LCA (2010-2020): Decade of Life Cycle Sustainability Analysis The LCSA framework is a framework for future LCA (see Figure 2.4). It broadens the scope of current LCA from mainly environmental impacts only to covering all three dimensions of sustainability (people, planet and prosperity). It also broadens the scope from predominantly product-related questions (product level) to questions related to sector (sector level) or even economy-wide levels (economy level). In addition, it deepens current LCA to also include other than just technological relations, e.g. physical relations (including limitations in available resources and land), economic and behavioral relations, etc. In addi- tion, as part of deepening, normative aspects such as discounting, weighting, A N OVERVIEW OF THE LIFE CYCLE ASSESSMENT METHOD 35 Life-Cycle Sustainability Analysis (LCSA) Goal and scope definition Environmental Modeling — Broadening the scope of indicators · t Eeon4my-wi<Je \Multi-region IOA/general\ equilibrium models/.. BOA/... 10A/partial equilibrium models/.. Product-oriented Process-LCA/EIO-LCA/ hybrid LCA LCC SLCA c Interpretation Figure 2.4 Trans-disciplinary integration framework for life cycle sustainability analysis (adapted from [83]). and weak versus strong sustainabilitycan be explicitly incorporated [82]. The term framework is used as LCSA, unlike LCA, is a trans-disciplinary integra- tion framework of models rather than a model in itself. LCSA works with a plethora of disciplinary models and guides selecting the proper ones, given a specific sustainability question. Structuring, selecting and making the pleth- ora of models practically available in relation to different types of life cycle sustainability questions is then the main challenge. Although this is fully com- patible with ISO's clause "there is no single method for conducting LCA," it is a significant deviation from LCA practice up until now. The broadening to economic and social impacts is also at variance with ISO's explicit restriction to environmental issues. There are three important differences compared to the ISO 14040 framework of Figure 2.1: • The merging of inventory analysis and impact assessment into one modeling phase (middle box). As has become clear during the last decade of academic work on agricultural production, climate change, impacts of land use, rebound and so on, it is difficult to 36 LIFE CYCLE ASSESSMENT HANDBOOK make a clear separation between behavior and technology, and between technosphere and ecosphere. The fuel needed to drive 1 km with a certain car depends on the car, the drive style, the road, other traffic, and the traffic policy. Actual impacts of a seem- ingly technological process such as transportation are thus linked to consumer behavior, policy-making, strategic investments, etc. Some endpoint models like Eco-Indicator 99 [33] and ReCiPe [84] include human adaptation scenarios in their endpoint models on climate change, but they do not include the environmental impli- cations of these adaption scenarios, such as the production of electricity to run additional air conditioners (as a consequence of global warming), or the production of additional sun blockers (as a consequence of ozone layer depletion). • The broadening of the object of analysis (vertical arrow). LCSA can be performed at three different levels: product, meso or economy. Products are thereby defined as in the ISO 14040 Standards and comprise any good or service. Product systems performing the same function(s) are compared, for example different options for milk packaging. Examples of methods and models for this level include process-LCA, EIO-LCA, hybrid LCA, life-cycle costing (LCC) and social LCA (SLCA). Meso refers to a level in-between product and economy-wide. It may include groups of related prod- ucts and technologies, baskets of commodities (e.g., the product folio of a company), a municipality, a household, etc. An example at this level might be the introduction of biomass as a major car fuel. Defining and finding appropriate methods and models for this level needs further research [85] but may for example include environmental input-output analysis (EIOA), input-output analysis (IOA) and partial equilibrium models. Economy-wide refers both to economies of states or other geographical/political entities, and eventually the world. An example question for this level might be the comparison of options for emerging technol- ogy domains, for example large-scale introduction of wind energy or solar cells as strategy for phasing out fossil energy, nanotech- nology and new communication services. Defining and finding appropriate methods and models for this level also needs further research but may for example include IOA [86] and multi-region IOA [87]. Obviously the three levels are not sharply defined, and there may be questions that fall somewhere in-between two levels. • The broadening of the scope of indicators (horizontal arrow). Analyses are made for at least one set of sustainability indicators (environmental, economic and/or social indicators). A distinction is made between LCA with just one set of sustainability indica- tors (environmental, economic or social), and LCSA comprised of performance indicators for all three (or at least two) pillars of sustainable development [86]. AN OVERVIEW OF THE LIFE CYCLE ASSESSMENT METHOD 37 The aspect of deepening is not shown in Figure 2.4. It mainly refers to the "Modeling" phase of LCSA. Deepening can be done in each box of the mod- eling phase. Consequential modeling is an example of deepening: it can be relevant and applied at each of the three levels of analysis and for each type of indicator. For a further discussion of the concept of deepening, we refer to [82]. References 1. ISO. Environmental management - life cycle assessment - requirements and guidelines (ISO 14044). International Organization for Standardization, Geneva, 2006. 2. Heijungs, R. and Suh, S. The computational structure of life cycle assessment. Kluwer Academic Publishers, Dordrecht, 2002. 3. Heijungs, R. and Kleijn, R. "Numerical approaches to life-cycle interpretation. Five exam- ples." International Journal of Life Cycle Assessment 6:3 (2001), 141-148. 4. Guinee, J.B., Heijungs, R., Huppes, G., Zamagni, A., Masoni, R, Buonamici, R., Ekvall, T., and Rydberg, T. "Life cycle assessment: past, present, and future." Environmental Science & Technology 45:1 (2011), 90-96. 5. Assies, J.A. "Introduction paper to SETAC-Europe workshop on environmental life cycle analysis of products." In Life-Cycle Assessment, Proceedings of a SETAC-Europe workshop on Environmental Life Cycle Assessment of Products, December 2-3 1991, Leiden; SETAC-Europe: Brussels, Belgium (1992). 6. Sundström, G. Investigation of energy requirements from raw material to garbage treatment for four Swedish beer and packaging alternatives. Malmö, Sweden, 1971. 7. Boustead, I. "Resource implications with particular reference to energy requirements for glass and plastic milk bottles." International Journal of Dairy Technology 1974,27 (3), 159-165. 8. Energy analysis workshop on methodology and conventions, August 25-30, 1974; International Federation of Institutes for Advanced Study (IFIAS): Guldsmedshyttan, Sweden, 1974. 9. Hunt, R.G., Franklin, W.E., Welch, R.O., Cross, J.A., and Woodal, A.E. Resource and envi- ronmental profile analysis of nine beverage container alternatives. U.S. Environmental Protection Agency: Washington, DC, 1974. 10. Studie Umwelt und Volkswirtschaft, Vergleich der Umweltbelastung von Behältern aus PVC, Glas, Blech und Karton; Basler & Hofman Ingenieure und Planer; Eidgenössisches Amt für Umweltschutz: Bern, Switzerland, 1974. 11. Ökobilanzen von Packstoffen. Schriftenreihe Umweltschutz no. 24; Bundesamt für Umwelt- schutz: Bern, Switzerland, 1984. 12. Guinee, J.B., Udo de Haes, H.A., and Huppes, G. "Quantitative life cycle assessment of products 1: Goal definition and inventory." /. Clean. Prod. 1993,1 (1), 3-13. 13. Product Life Assessments: Policy issues and implications; Summary of a Forum on May 14,1990; World Wildlife Fund and The Conservation Foundation: Washington, DC, 1990. 14. Fava, J.A., Denison, R., Jones, B., Curran, M.A., Vigon, B., Selke, S., and Barnum, J., Eds. A Technical Framework for Life-Cycle Assessments; Workshop Report Society of Environmental Toxicology and Chemistry; SETAC: Washington, DC, 1991. 15. Smet, B. de, Ed. Life-cycle analysis for packaging environmental assessment; Proceedings of the specialised workshop, 24-25 September 1990, Leuven. Procter & Gamble Technical Center: Strombeek-Bever, Belgium, 1990. 16. Life-Cycle Assessment; Proceedings of a SETAC-Europe workshop on Environmental Life Cycle Assessment of Products December 2-3 1991, Leiden; SETAC-Europe: Brussels, Belgium, 1992. 17. Fava, J.A., Consoli, F, Denison, R., Dickson, K., Mohin, T, and Vigon, B., Eds. A Concept- ual Framework for Life-Cycle Impact Assessment; Society of Environmental Toxicology and Chemistry and SETAC Foundation for Environmental Education, Inc. Workshop Report; SETAC: Pensacola, Florida, 1993. 38 LIFE CYCLE ASSESSMENT HANDBOOK 18. Huppes, G. andSchneider, R, Eds. Proceedings of the European Workshop on Allocation in LCA under the Auspices of SETAC-Europe, February 24-25,1994, Leiden; SETAC-Europe: Brussels, Belgium, 1994. 19. Umweltprofile von Packstoffen und Packmitteln: Methode; Fraunhofer-Institut für Leben- smitteltechnologie und Verackung: München; Gesellschaft für Verpackungsmarktforschung Wiesbaden und Institut für Energie- und Umweltforschung Heidelberg: Germany, 1991. 20. Grieshammer, R., Schmincke, E., Fendler, R., Geiler, N., and Lütge, E. Entwicklung eines Verfahrens zur ökologischen Beurteilung und zum Vergleich verschiedener Wasch- und Reinigungsmittel; Band 1 und 2. Umweltbundesamt: Berlin, Germany, 1991. 21. Product Life Cycle Assessment - Principles and Methodology; Nord 1992: 9, Nordic Council of Ministers: Copenhagen, Denmark, 1992. 22. Heijungs, R., Guinee, J.B., Huppes, G., Lankreijer, R.M., Udo de Haes, H.A., Wegener Sleeswijk, A., Ansems, A.M.M., Eggeis, P.G., Duin, R. van, and Goede, H.P. de. Environmental life cycle assessment of products. Guide & Backgrounds - October 1992; Centre of Environmental Science, Leiden University: Leiden, The Netherlands, 1992. 23. Vigon, B.W., Tolle, D.A., Cornaby, B.W., Latham, H.C., Harrison, C.L., Boguski, T.L., Hunt, R.G., and Sellers, J.D. Life-Cycle Assessment: Inventory Guidelines and Principles; EPA/600/R-92/245; Environmental Protection Agency: Washington, DC, 1993. 24. Lindfors, L.-G. , Christiansen, K., Hoffman, L., Virtanen, Y, Juntilla, V., Hanssen, O.J., Ronning, A., Ekvall, T., and Finnveden, G. Nordic Guidelines on Life-Cycle Assessment, Nord 1995:20; Nordic Council of Ministers: Copenhagen, Denmark, 1995. 25. Curran, M.A. Environmental Life-Cycle Assessment; McGraw-Hill: New York, 1996. 26. Hauschild, M. and Wenzel, H. Environmental Assessment of products. Volume 1: Methodology, tools and case studies in product development - Volume 2: Scientific background; Chapman & Hall: London, U.K., 1998. 27. Consoli, R, Allen, D., Boustead, I., Oude, N. de, Fava, J., Franklin, W, Quay, B., Parrish, R., Perriman, R., Postlethwaite, D., Seguin, J., and Vigon, B., Eds. Guidelines for Life-Cycle Assessment: A ^Code of Practice/ 1st ed.; SETAC-Europe: Brussels, Belgium, 1993. 28. Directive 94/62/EC, OJ L 365, 31.12.1994, pplO-23; http://eur-lex.europa.eu/LexUriServ/ LexUriServ.do?uri=CELEX: 31994L0062: EN: HTML. 29. Hunkeler, D., Yasui, I., and Yamamoto, R. "LCA in Japan: policy and progress." Int J Life Cycle Assess 1998,3 (3), 124-130. 30. Owens, J.W. "Life-cycle assessment: Constraints on moving from inventory to impact assess- ment." Journal of Industrial Ecology 2000,4 (3), 11-33. 31. Arnold, F.S. "Why environmental life cycle assessment doesn't work." Journal of Environmental Law & Practice 1995,2 (5), 4-14. 32. Hofstetter, P. Perspectives in life cycle Impact assessment: A structured approach to combine models of the technosphere, ecosphere and valuesphere. Kluwer Academic Publishers: Dordrecht, the Netherlands, 1998. 33. Goedkoop, M. and Spriensma, R. The Eco-indicator 99 - A damage oriented method for life cycle Impact assessment; PRe Consultants: Amersfoort, the Netherlands, 1999. 34. Hauschild, M.Z., Huijbregts, M., Jolliet, O., Macleod, M., Margni, M., Meent, D. van de, Rosenbaum, R.K., and McKone, T.E. "Building a Model Based on Scientific Consensus for Life Cycle Impact Assessment of Chemicals: The Search for Harmony and Parsimony." Environ. Sei. Technol. 2008,42 (19), 7032-7037. 35. Rosenbaum, R.K., Bachmann, T.M., Gold, R.S., Huijbregts, M.A.J., Jolliet, O., Juraske, R., Köhler, A., Larsen, H.F., MacLeod, M., Margni, M., McKone, T.E., Payet, J., Schuhmacher, M., Meent, D. van de, and Hauschild, M.Z. "USEtox—the UNEP-SETAC toxicity model: recom- mended characterisation factors for human toxicity and freshwater ecotoxicity in life cycle impact assessment." Int J Life Cycle Assess 2008,13 (7), 532-546. 36. Guinee, J.B. and Heijungs, R. "A proposal for the classification of toxic substances within the framework of Life Cycle Assessment of Products." Chemosphere 1993, 26 (10), 1925-1944. A N OVERVIEW OF THE LIFE CYCLE ASSESSMENT METHOD 39 37. Weidema, B.P., Frees, N., and Nielsen, P. "Marginal production technologies for life cycle inventories." Int J Life Cycle Assess 1999,4 (1), 48-56. 38. Weidema, B. "Avoiding Co-Product Allocation in Life-Cycle Assessment." Journal of Industrial Ecology 2000,4 (3), 11-33. 39. Ekvall, Τ. "A market-based approach to allocation at open-loop recycling." Resources, Conservation and Recycling 2000,29 (1-2), 93-111. 40. UN Environment Programme Life Cycle Initiative website; http://lcinitiative.unep.fr/. 41. "Integrated Product Policy - Building on Environmental Life-Cycle Thinking." Commission of the European Communities, COM(2003) 302 final, Brussels, Belgium, 2003; http://eur- lex.europa.eu/LexUriServ/site/en/com/2003/com2003_0302en01.pdf. 42. "Thematic strategy on the sustainable use of natural resources." Commission of the European Communities, COM(2005) 670 final, Brussels, Belgium, 2005; http://eur-lex.europa.eu/ LexUriServ/LexUriServ.do?uri-COM: 2005: 0670: FIN: EN: PDF. 43. "Taking sustainable use of resources forward: A Thematic Strategy on the prevention and recycling of waste." Commission of the European Communities, COM(2005) 666 final, Brussels, Belgium, 2005; http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=COM: 2005: 0666: FIN: EN: PDF. 44. European Commission - Joint Research Centre Life Cycle Thinking and Assessment Website; http: / /let. jrc.ec.europa.eu / . 45. U.S. Environmental Protection Agency Life-Cycle Assessment Research Website; http:// www.epa.gov / nrmrl / lcaccess / . 46. ALCAS Australian Life Cycle Assessment Society Website; http://www.alcas.asn.au/. 47. ACLA American Center for Life Cycle Assessment Website; http://www.lcacenter.org/. 48. Thai LCA Network Website; http://www.thailca.net/. 49. Energy independence and security act of 2007. Public Law 110-140, 2007; http://frweb gate.access.gpo.gov/cgi-bin/getdoc.cgi?dbname)110_cong_public_law&docid)f:publl40.110. pdf. 50. U.S. Environmental Protection Agency: Lifecycle Analysis of Greenhouse Gas Emissions from Renewable Fuels Website; http://www.epa.gov/otaq/renewablefuels/420f09024. htm. 51. International Standardisation of Carbon Footprinting Website; http://www.lcacenter.org/ LCA9/special/Carbon-footprint.html. 52. Matthews, S.H., Hendrickson, C, and Weber, C.L. "The Importance of Carbon Footprint Estimation Boundaries." Environ. Sei. Technol. 2008,42 (16), 5839-5842. 53. Guinee, J.B., Heijungs, R., and Voet, E. van der. "A greenhouse gas indicator for bio-energy: some theoretical issues with practical implications." Int J Life Cycle Assess 2009, 14 (4), 328-339. 54. Voet, E. van der, Lifset, R.J., and Luo, L. "Life Cycle Assessment of biofuels, convergence and divergence." Accepted for publication in Biofuels 2010,1 (3), 435-449. 55. Sheehan, J.J. "Biofuels and the conundrum of sustainability." Current Opinion in Biotechnology 2009,20 (3), 318-324. 56. Heijungs, R., Huppes, G., and Guinee, J.B. A scientific framework for LCA; Deliverable 15 of the CALCAS project, 2009. Available at http://www.estis.net/sites/calcas/. 57. Schmidt, J.H. "Comparative life cycle assessment of rapeseed oil and palm oil." Int J Life Cycle Assess 2010,15 (2), 183-197. 58. "Intergovernmental Panel on Climate Change: Special report on Emissions Scenarios 2000." Available at http://www.grida.no/publications/other/ipcc_sr/?src=/climate/ ipec/emission/. 59. Graedel, T.E. and van der Voet, E., Eds. Linkages of Sustainability', Striingmann Forum Reports; The MIT Press: Cambridge Massachusetts, 2008. 60. Finnveden, G., Hauschild, M.Z., Ekvall, T, Guinee, J., Heijungs, R., Hellweg, S., Koehler, A., Pennington, D., and Suh, S. "Recent developments in life cycle assessment." Journal of Environmental Management 2009,91 (1), 1-21. 40 LIFECYCLE ASSESSMENT HANDBOOK 61. Zamagni, A., Buttol, P., Porta, PL., Buonamici, R., Masoni, P., Guinee, J.B., Heijungs, R., Ekvall, T., Bersani, R., Bienkowska, A., and Pretato, U. Critical review of the current research needs and limitations related to ISO-LCA practice; Deliverable 7 of the CALCAS project, 2008. Available at http://www.estis.net/sites/calcas/. 62. Pehnt, M. "Dynamic life cycle assessment (LCA) of renewable energy technologies." Renew. Energy 2006,31 (1), 55-71. 63. Björk, H. and Rasmuson, A. "A method for life cycle assessment environmental optimisation of a dynamic process exemplified by an analysis of an energy system with a superheated steam dryer integrated in a local district heat and power plant." Chemical Engineering Journal 2002, 87 (3), 381-394. 64. Levasseur, A., Lesage, P., Margni, M., Deschenes, L., and Samson, R. "Considering Time in LCA: Dynamic LCA and Its Application to Global Warming Impact Assessments." Environ. Sei Technol. 2010,44 (8), 3169-3174. 65. Kendall, A., Chang, B., and Sharpe, B. "Accounting for Time-Dependent Effects in Biofuel Life Cycle Greenhouse Gas Emissions Calculations." Environ. Sei. Technol. 2009, 43 (18), 7142-7147. 66. Assies, J.A. "A risk-based approach to life-cycle impact assessment." Journal of Hazardous Materials 1998,61 (1), 23-29. 67. Nishioka, Y., Levy, J.I., and Norris, G.A. "Integrating air pollution, climate change, and eco- nomics in a risk-based life-cycle analysis: a case study of residential insulation." Human and Ecological Risk Assessment 2006,12 (3), 552-571. 68. Saouter, Ε. and Feijtel, T.C.J. "Use of life cycle analysis and environmental risk assess- ment in an integrated product assessment. Environmental Strategies." In Risk Assessment and Life Cycle Assessment, TemaNord 2000: 545; Hauschild, M., Olsen, S.I., Poll, C, and Bro- Rasmussen, R, Eds.; Nordic Council of Ministers: Copenhagen, Denmark, 2000. 69. Sonnemann, G., Castells, R, and Schuhmacher, M. Integrated life-cycle and risk assessment for industrial processes; Lewis Publishers: Boca Raton, Florida, 2004. 70. Suh, S., Lenzen, M., Treloar, G.J., Hondo, H., Horvath, A., Huppes, G., Jolliet, O., Klann, U., Krewitt, W., Moriguchi, Y, Munksgaard, J., and Norris, G. "System boundary selection in life-cycle inventories using hybrid approaches." Environ. Sei. Technol. 2004,38 (3), 657-664. 71. Hendrickson, C, Lave, L.B., and Matthews, H.S. Environmental life cycle assessment of goods and services - An input-output approach; Resources for the Future: Washington, DC, 2006. 72. Heijungs, R., de Koning, A., Suh, S. and Huppes, G. "Toward an information tool for inte- grated product policy: requirements for data and computation." Journal of Industrial Ecology 2006,10 (3), 147-158. 73. Hunkeler, D., Lichtenvort, K., and Rebitzer, G., Eds. Environmental Life Cycle Costing. CRC Press: New York, 2008. 74. Sherif, Y.S. and Kolarik, W.J. "Life cycle costing: concept and practice." OMEGA: The International Journal of Management Service 1981,9 (3), 287-296. 75. Benoit, C. and Mazijn, B., Eds. Guidelines for Social Life Cycle Assessment of Products; UNEP/ SETAC Life Cycle Initiative: Paris, 2009. Available at http://www.estis.net/includes/file. asp?site=lcinit&file=524CEB61-779C-4610-8D5B-8D3B6B336463. 76. Settanni, E. The need for a computational structure of LCC. Int J Life Cycle Assess 2008,13 (7), 526-531. 77. Huppes, G., Rooijen, M. van, Kleijn, R., Heijungs, R., Koning, A. de, and Oers, L. van. Life Cycle Costing and the Environment. Institute of Environmental Sciences (CML), Universiteit Leiden: Leiden, the Netherlands, 2004. http://www.rivm.nl/milieuportaal/images/ Report%20LCC%20April%20%202004%20final.pdf. 78. Guinee, J.B., Ed. Handbook on Life Cycle Assessment: Operational Guide to the ISO Standards; Springer - Kluwer Academic Publishers: Dordrecht, the Netherlands, 2002. 79. Baumann, H. and Tillman, A.-M. The Hitch hiker's Guide to LCA - An orientation in life cycle assessment methodology and application; Studentlitteratur: Lund, Sweden, 2004. 80. ILCD Handbook: General guide for life cycle assessment - provisions and action steps; European Commission, JRC-IES: Ispra, Italy, 2010. Available at http://lct.jrc.ec.europa. A N OVERVIEW OF THE LIFE CYCLE ASSESSMENT METHOD 41 eu/pdf-directory/ILCD-Handbook-General-guide-for-LCA-PROVISIONS-online- 12March2010.pdf. 81. CALCAS Co-ordination Action for innovation in Life-Cycle Analysis for Sustainability Website; http: / / www.calcasproject.net/. 82. Heijungs, R., Huppes, G., and Guinee, J.B. "Life cycle assessment and sustainability analysis of products, materials and technologies. Towards a scientific framework for sustainability life cycle analysis." Polymer Degradation and Stability 2010,95 (3), 422-428. 83. Zamagni, A., Buttol, R, Buonamici, R., Masoni, P., Guinee, J.B., Huppes, G., Heijungs, R., Voet, E. van der, Ekvall, T., and Rydberg, T. Blue Paper on Life Cycle Sustainability Analysis; Deliverable 20 of the CALCAS project, 2009. Available at http://www.estis.net/sites/ calcas/. 84. Goedkoop, M., Heijungs, R., Huijbregts, M.A.J., De Schryver, A.M., Struijs, J., Zelm, J, R. van. 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, 2009. Available at http://www.lcia-recipe.net/. 85. Guinee, J.B., Huppes, G., Heijungs, R., and Voet, Ε. van der. Research strategy, programmes and exemplary projects on life cycle sustainability analysis (LCSA); Deliverable 22 of the CALCAS project, 2009. Available at http://www.estis.net/sites/calcas/. 86. University of Sydney: Balancing Act - A Triple Bottom Line Analysis of the Australian Economy. Available at: http://www.cse.csiro.au/publications/2005/balancingactl.pdf. 87. Sonis, M., Oosterhaven, J., and Hewings, G.J.D. "Spatial Economic Structure and Structural Changes in the EC: Feedback Loop Input-Output Analysis." Economic Systems Research 1993, 5 (2), 173-184. 3 Life Cycle Inventory Modeling in Practice Beverly Sauer Franklin Associates, A Division of ERG, KS, USA Abstract This chapter focuses on the first two phases of LCA, goal and scope definition and the life cycle inventory (LCI) analysis, which provide the necessary foundation for the impact assessment and interpretation stages of LCA. A clearly defined scope is impor- tant to identify the intended audience and set the boundaries of the study. The author discusses practical considerations and approaches for issues commonly encountered in life cycle inventory modeling (the reader is referred to the ISO 14040 and 14044 stan- dards for a complete description of LCA methodology). Several important method- ological issues for creating LCI are discussed, including areas where different choices can be made depending on the specific systems being studied. This chapter also covers issues that are the focus of increasing interest in the environmental community, such as water use and carbon tracking. As LCA evolves, practitioners should stay abreast of methodological developments, and follow accepted LCI methodology and best prac- tices when conducting the scoping and inventory stages. Keywords: Life cycle assessment, life cycle inventory, goal definition, scope 3.1 Introduction A Life Cycle Assessment (LCA) consists of four phases: • Goal and scope definition, • Inventory analysis (LCI), • Impact assessment (LCIA), and • Interpretation. The life cycle inventory (LCI) provides the foundation of the life cycle assessment (LCA). The inventory of flows to and from the natural environ- ment serves as the basis for the impact assessment and interpretation phases of the LCA, so it is critically important that the inventory be methodologically sound, complete, and unbiased. International Standards ISO 14040 and 14044 provide detailed guidance on the key aspects of life cycle inventorymethodology [1,2]. This chapter will Mary Ann Curran (ed.) Life Cycle Assessment Handbook: A Guide for Environmentally Sustainable Products, (43-66) © 2012 Scrivener Publishing LLC 43 44 LIFE CYCLE ASSESSMENT HANDBOOK not attempt to cover all the content of the standards but will discuss practical considerations and approaches for issues commonly encountered in life cycle inventory modeling. In some areas, the ISO standards allow for some flex- ibility, allowing the practitioner to make methodological choices that are most relevant and appropriate based on the characteristics of the specific systems being analyzed. This chapter focuses on the first two phases of LCA, the goal and scope defi- nition and the life cycle inventory analysis phase. 3.2 Study Goal Defining the goal of an LCA includes defining the application(s) to be studied, the reasons for carrying out the study, the audience to whom the results are to be communicated, and whether the results will be used as the basis for public comparative assertions [1]. Because the scope of the study, including the study boundaries and the level of detail, depends upon the goal of the study, the first step in the LCI process is to clearly define the goal. The goal of a study may be relatively simple and straightforward, for exam- ple, to assess the energy and greenhouse gas impacts associated with produc- tion of a single product, for internal use by the producer as a benchmark for evaluating future process improvements or design changes. In other cases the goal can lead to a very complex analysis, for example, a study with the goal of make public comparative claims about environmental performance of several competing products with variations in functional properties. The definition of the goal and intended use will guide the practitioner in set- ting the scope and boundaries for the analysis, including the need for critical review and the type of critical review required. LCAs can be conducted on a single system, but most are comparative. Study results may be intended for internal use or for sharing with external parties. Examples of types of LCAs and goals include the following: • Single System - Internal Use of Results • Analyze current product to identify opportunities for reducing environmental impact • Establish product baseline against which to measure future improvements • Single System - External Use of Results • Environmental product declaration (e.g., to share with custom- ers who request information about environmental metrics for product) • Comparative Analysis - Internal Use of Results • Compare alternative design options for company's own prod- uct or packaging • Compare new concept design with alternatives already in the marketplace to make a business development decision LIFE CYCLE INVENTORY MODELING IN PRACTICE 45 • Comparative Analysis - External Use of Results • Provide science-based defense to public concerns or criticisms of a product's environmental performance compared to alter- natives, including proposed legislation or bans • Use LCA results as the basis for marketing statements compar- ing a company's product with competing products It is recommended, and in some cases required, that studies intended for external use undergo a critical review. For environmental product declara- tions prepared under product category rules, the program operator generally defines the peer review process. Critical review by a panel of interested parties is required for LCAs intended to be used as the basis for public comparative claims about the environmental superiority of one system over another [2]. 3.3 Scope Once the study goal has been defined, scoping decisions can be made. The breadth and level of detail of the study must be sufficient to ensure that the stated goals have been addressed. Key aspects of the scoping phase include defining the product system(s) to be studied, the functional unit, the system boundaries (in terms of life cycle stages as well as time and geographic bound- aries), methodological issues such as allocation procedures, impact categories to be included, impact assessment method, and need for critical review. As the project proceeds, some of the initial goals (such as goals for age of data used, industry coverage, etc.) may not be met, as they may be subject to data avail- ability and practical constraints on project budget and timing. It is important to document not only the intended goals but also the limitations that were encountered in conducting the analysis and assumptions or estimates that were made. 3.3.1 Functional Unit As defined in ISO 14040, "The functional unit defines the quantification of the identified functions (performance characteristics) of the product. The primary purpose of a functional unit is to provide a reference to which the inputs and outputs are related. This reference is necessary to ensure comparability of LCA results" [1]. If the goal of the project is to develop an environmental profile for an indi- vidual material, then the functional unit can be very simple, e.g., output of a kilogram (kg) of material or a megajoule (MJ) of a fuel. However, defining the functional unit for comparative analyses of products can become quite complicated. In a comparative analysis, the functional unit must take into account differences in the properties of the product, such as strength or durability, 46 LIFE CYCLE ASSESSMENT HANDBOOK or differences in the use phase of the product. If the properties and per- formance of each of the systems analyzed is the same, the systems can be compared on a one-to-one basis. However, there are often differences in the systems that must be taken into consideration when defining the functional unit and scope of the analysis. Some differences are relatively simple to accommodate. For example, consider a comparative analysis of two multi-serving bever- age containers with differences in capacity. Since the function of each container is to deliver beverage to the consumer, an appropriate basis of comparison would be delivery of an equivalent number of fluid ounces or servings. To deliver the equivalent volume of beverage, fewer of the larger volume contain- ers would be required. In the preceding example, the beverage inside the containers was the same, and only the container sizes were different. A related example is products that deliver the same functional service with different volumes of packag- ing due to differences in the form of product. Examples include concentrated detergent compared to pre-diluted detergent, condensed soup compared to ready-to-serve soup, and pre-ground coffee compared to coffee beans. In each case, the condensed or more compact version of the product requires less packaging to deliver the same number of servings or uses. It is worth not- ing, however, that differences in concentration may introduce differences in the product use phase. Preparation of condensed soup, for example, requires the consumer to put the soup into a larger container for dilution and heating prior to serving. This results in additional container washing burdens for the condensed soup compared to ready-to-serve soup packaged in a microwave- able container that serves a dual function of packaging and serving the soup. Purchasing pre-ground coffee rather than coffee beans eliminates the need for the consumer to grind the coffee prior to use. A comparison of pre-ground cof- fee with coffee beans would thus need to include not only the differences in packaging requirements but also the differences in commercial grinding and home grinding operations. In other cases, data may not be available to precisely quantify functional differences between systems, or there may be differences due to consumer behavior rather than the inherent properties of the product, as in the following examples: Example 1. Functional equivalence varies from useto use within the defined application. Example system: Disposable plates. The function of a disposable plate is to hold a quantity of food. Some plates are very lightweight and will only support a few ounces of food, while other plates are heavier and sturdier and can hold much greater loads. For a light duty use (such as a piece of cake or some snack crackers), a single plate of either type will be sufficient to support the load, and the plates will have a one- to-one functional equivalence. However, if the intended application is a full meal that includes heavier items, then two or more lightweight plates may be required to hold the same amount of food that can be placed on a single heavyweight plate. In cases where functional equivalence can vary within LIFE CYCLE INVENTORY MODELING IN PRACTICE 47 an expected range of use applications, it is advisable to structure the analy- sis so that the products can be compared on single or multiple-unit bases, depending on use variations. Example 2. Consumer behavior overrides true functional equivalence. Example system: Cell phone. Company A manufactures a cell phone that will last for 6 years of normal use, while Company B's phone is designed to last only 3 years. On a true functional equivalence basis, two company B phones would be required to provide the same years of use as one company A phone. However, if the average consumer replaces their cell phone after 3 years in order to upgrade to the latest technology, then both phones would be replaced after 3 years, and the company A phone would not remain in use long enough to show a durability benefit over the company B phone. In some cases, the characteristics of different products make it impossible to compare them on a completely equivalent basis. For example, carpet, hard- wood, and ceramic tile all provide the same basic function of covering a given area of interior floor surface. The life cycle assessment can take into account the differences in the composition, manufacture, useful lifetime, care and maintenance during use, and end-of-life management for the different floor coverings. However, a square meter of carpeted floor still has very different properties and qualities than a square meter of hard surface floor covering. A consumer making a decision about the type of floor covering to use in a kitchen or bedroom would not consider carpet, hardwood, and tile to be com- pletely equivalent flooring options. An important consideration when comparing packaging systems is protec- tive performance. The function of packaging is to deliver undamaged product to a consumer. The environmental impacts associated with production of the product inside the package are typically much greater than the impacts asso- ciated with the package. Therefore, if one type of package is more effective in protecting product than another package, the differences associated with product damage can outweigh any environmental impacts associated with dif- ferences in the packaging. For example, a variety of types and quantities of packaging shapes can be used to support and protect a computer inside a cor- rugated box. The environmental impacts for producing the computer are far greater than the impacts for producing the packaging shapes. Suppose pack- aging system 1 has higher production impacts than packaging system 2, but system 1 provides better damage protection than system 2. If use of system 1 prevents damage to even one more computer than system 2, then packaging system 1 is a better choice. 3.3.2 Boundaries System boundaries must be defined in terms of the life cycle stages to be included in the analysis, the geographic and time boundaries of the analysis, and the flows and impact categories to be included. Life Cycle Stages. Defining the life cycle stages that will be included in the scope of the study is a critically important step. Figure 3.1 provides a basic 48 LIFE CYCLE ASSESSMENT HANDBOOK illustration of the stages in a full life cycle inventory, beginning with raw mate- rial extraction and continuing through end-of-life management of the finished product. For each stage, the inventory quantifies the incoming flows from nature and from the technosphere, including materials and energy, as well as the outputs of useful products, co-products, and wastes, including solid wastes and emissions released to air and water. Transportation between life cycle stages is also included. Figure 3.2 illustrates the stages for the life cycle of a PET bottle, such as those used for bottled water and soft drinks. (Note that this diagram focuses on the bottle and does not show production of the bottle cap, label, product in the bottle, or packaging used for shipping filled bottles. A full life cycle for a bottled product would include all the additional components.) Life cycle stages often include complex networks of unit processes. For example, the steps required for cradle-to-resin production of virgin PET resin (shown as the first two highlighted blocks on Figure 3.2) expands to the net- work of processes shown Figure 3.3. The scoping process must capture all stages and operations that are needed for the functional equivalence basis that has been selected for the analysis. Energy 1 Energy 1 Product use or consumption Wastes r Wastes Recycle Figure 3.1 General life cycle flow diagram. Wastes Energy 1 PET resin production injection stretch blow molding Reclaim Figure 3.2 Life cycle of a PET bottle. (light gray highlighted stages expanded to show unit processes in Figure 3.3) LIFE CYCLE INVENTORY MODELING IN PRACTICE 49 Natural g production Natural gas Crude oil production Methanol manufacture Distillation/ desalting Hydrotreating Ethylen« manufacture Mixed xylenes Acetic acid manufacture Carbon monoxide manufacture ture I H TPA manufacture (1) DMT manufacture Methanol recycle w PTA manufacture Melt phase and solid state PET polymerization from DMT h t Ethylene oxide manufacture Paraxylene extraction I I Ethylene | H glycol 1 I manufacture I Melt phase and solid state PET polymerization from PTA h Figure 3.3 Cradle-to-resin production of virgin PET. Before making the decision to exclude a life cycle stage, the implications must be carefully considered, as illustrated in the following example. A company wishes to conduct an LCA to compare a new flexible composite packaging option with a current rigid mono-material package. Both packages hold the same quantity of product. The flexible packaging option is lighter in weight than the current package, but the current package is recyclable and the new composite package is not. Based on these differences in the packages, the decision is made to include the following life cycle stages in the analy- sis: raw material extraction through container manufacturing, and end-of-life management. This choice of boundaries excludes several potentially significant differences from the analysis. First, it is likely that the rigid container will have higher burdens for empty container transport to the filler. The flexible package can be shipped flat, so that shipping is much more efficient compared to the rigid packages, which occupy a large amount of space per empty container. Second, even though the containers hold the same volume of product, there was no consideration of potential differences in the filling operation for the two types of containers. A third overlooked area was transport of the packages after fill- ing. Although the rigid containers are heavier than the flexible packages, the flexible packages may require more secondary packaging to stabilize and pro- tect the filled packages during shipment. For example, the rigid containers might be packed in corrugated trays, while the flexible package might require the extra support and protection of a fully enclosed box, addingsecondary packaging material and weight. Systems providing functional equivalence may have very different life cycle operations. An example is single-use and reusable container systems. 50 LIFE CYCLE ASSESSMENT HANDBOOK Example: Shipping containers. The function of a shipping container is to deliver an undamaged quantity of goods. Therefore, for con- tainers that have equivalent capacity and provide equivalent prod- uct protection, an appropriate functional unit might be defined as 1,000 container shipments. If the containers being compared are a single-trip container and a reusable container, it is important to con- sider all the differences involved in making an equivalent number of shipments in each type of container. For the single-use container, each time a shipment is made, a con- tainer must be manufactured and shipped to the filler. At the end user destination, the emptied container must be managed by recy- cling, landfill disposal, or combustion. For the reusable container, the number of containers that must be manufactured for a given number of shipments depends on many factors including the percentage of containers that are lost or sto- len during use, damage rates, fate of damaged containers (repaired, recycled, or disposed), and lifetime uses for containers that remain in circulation. Although fewer reusable containers must be manufac- tured for a given number of shipments, there are additional impacts associated with reuse that are not required for disposable contain- ers. Reusable containers must be backhauled to the filler or to an inspection point. Backhauling may be done on a truck that is already returning from a retail store to a distribution center, or a special back- haul trip may be required. Reusable containers may require cleaning or reconditioning before they can be returned to use, which adds environmental burdens for materials and energy used in the clean- ing process. Depending on where the cleaning takes place, additional transportation may also be required. Other differences between systems may involve a combination of physical differences and consumer behavior. Consider the example of two ice cream cartons that hold an equivalent volume of product. Carton A is cylindrical with a lift-off lid, and Carton B is rectangular, with a paperboard flap closure. Carton B can be packed more compactly in a store freezer, so that less shelf space is required for ice cream packaged in Carton B compared to the same quantity of ice cream packaged in Carton A. Because Carton A occupies more freezer space per unit volume of ice cream, Carton A is allocated a larger share of the daily energy requirements for operating the store freezer. However, consumers tend to prefer Carton A's removable lid design over Carton B's paperboard flap closure. If consumer preference for the Carton A design translates into fas- ter sales compared to Carton B, the reduced time in the retail freezer for Carton A may offset its additional freezer shelf space requirement. When consumer behavior is involved, it is advisable to conduct a sensitivity analysis unless data are available to reliably characterize actual consumer behavior. Geographic Boundaries. The geographic boundaries of the system influence factors such as raw material sourcing, technology used, electricity grids, and LIFE CYCLE INVENTORY MODELING IN PRACTICE 51 transportation distances. Some materials have international raw material sup- ply chains that are dependent on the geographic distribution of ores and other natural resources. Other processes may be specifically located to take advan- tage of material or energy supplies. For example, electricity-intensive alumina smelting operations are sited to take advantage of hydropower. Modeling the aluminum supply chain using average U.S. grid electricity for all the processes would result in a large overstatement of carbon dioxide emissions associated with smelting. Other operations tend to be widely distributed, for example, converting operations such as plastic molding facilities. The locations of raw materials and subsequent processing operations should be taken into account when modeling transportation distances between life cycle stages or choosing elec- tricity grids for process energy at different life cycle stages. Regional issues are especially influential when evaluating agricultural sys- tems. Regional differences in climate, soil composition, rainfall, etc. can result in significant regional variations in crop yields, irrigation requirements, and applications and runoff of agricultural chemicals. Geographic boundaries also influence end-of-life scenarios. Recycling rates for products may vary widely from country to country or from region to region within a country, influenced by legislation, consumer behavior, and access to recycling programs. For example, beverage container recycling rates in U.S. states with deposit laws are higher than recycling rates for states without deposit laws. Landfill space is limited in some regions, while other regions may not have access to waste-to-energy combustion facilities. For material that is landfilled at end of life, local conditions such as temperature, moisture, etc. will affect the degree and rate at which biodegradable materials decompose (or fail to decompose) in a local landfill. Time Boundaries. Similar to geographic boundaries, time boundaries can influence the accuracy and relevance of study results. Systems should be modeled using the technology or mix of technologies that is relevant for current production in the region where processes take place. End-of-life management practices also change over time in response to increasing environmental awareness of consum- ers, legislative measures, and access to different recycling and disposal options. For long-lived products that have significant use-phase energy impacts, such as building insulation, greenhouse gas savings over time may be influenced by shifts in the energy supply (e.g., from fossil fuel-based electricity generation to more wind and solar electricity, or more use of biomass fuels), even if the prod- uct's properties do not change over time. In other words, an insulation product may provide consistent insulating performance over a 50-year lifetime, but annual greenhouse gas savings over time could become smaller as the energy supply becomes less carbon-intensive. Impact Categories. When scoping a project, it is important to define the impact categories that will be included in the results, as this influences the data collection requirements. A study scoped as a life cycle carbon footprint analy- sis will have very different data requirements than a study that includes a full set of life cycle impact indicators. 52 LIFE CYCLE ASSESSMENT HANDBOOK When gathering primary data or developing a unit process data set from secondary data sources, it is generally advisable to collect as much data as pos- sible at the greatest level of detail possible. For example, a unit process data developed for a carbon footprint study may later be needed for use in a full impact assessment study. If only greenhouse gas-related flows were included in the original data set, then the data set will not be usable for analyses that include additional impact indicators. Figure 3.4 provides an example of a form that can be used for collecting life cycle inventory data. More examples can be found in ISO 14044 Annex A: Company name Contact name Facility location TIME COVERAGE Start End Contact telephone Contact e-mail Month Year COMPANY COVERAGE Percent of company output of this material covered by this facility TECHNOLOGY REPRESENTED (brief description) INDUSTRY COVERAGE Percent of industry output of this material covered by this facility REPORTING BASIS (basb on which all data In the form are reported; eg , 1000 lb of output, specified number of product units, etc.) Product OutputBasis (with units) _ _ _ _ _ _ _ _ _ _ _ _ _ _ ^ _ _ _ ^ « » _ » _ _ _ _ _ _ _ _ « _ » » « _ . MATERIAL INPUTS TO PROCESS List all raw material inputs to the process, normalized to the product output basis specified above. Include incoming transport information (miles and mode: truck, rail, ship, etc.) Units (lb, Material name Amount kg, etc.) Distance Mode Distance Mode Name Amount Source (river, lake, ocean, etc.) Amount Units (lb, kg, etc.) Units (gal, liters, etc.) Add lines as needed USEFUL OUTPUTS Primary product Coproducts Recovered energy/heat Recycled scrap Add lines as needed WATER USAGE Process water Intake Output Cooling water Intake Output ENERGY USE (direct process energy and fuels used to produce process steam) Units (kWh, gal, Amount liters, etc.) Electricity (grid) Electricity (generated onsite) mmmmmmmmmmmmm _ Natural gas Distillate oil Residual oil Coal Add lines as needed Fuel for onsite generation LIFE CYCLE INVENTORY MODELING IN PRACTICE 53 SOLID WASTES Scrap that is recycled on or ofifcite should be reported in USEFUL OUTPUTS section. Units (lb, Type of waste Amount kg, etc.) Process wastes Scrap not recycled _ _ « _ Wastes collected in emission control devices Packaging wastes not recycled Add lines as needed EMISSIONS Report amount of each substance released to environment AFTER on-site emission controls or wastewater treatment. Disposition (landfill, burned, burned with energy recovery, etc.) Emfesiofls to Air Name Amount Units (lb, kg, etc.) Indicate emissions included Fuel Process combustion Add lines as needed Cmissioa· to Water Name Units (lb, kg, etc.) Add lines as needed PACKAGING FOR OUTGOING SHIPMENT Quantity Product shipped in units of «««■«««■■. Indicate emissions included Fuel Process combustion Receiving body of water (river, lake, municipal sewer, etc.) Units of product (items, lb, kg, etc.) per shipping unit List material used per shipping unit of product (corrugated boxes, plastic film sleeves, etc.) Units (lb, Packaging Material Amount kg, etc.) DATA COLLECTION METHOD AND SOURCES For each flow, check the appropriate box that describes the data sources. Provide additional descriptions as needed. Materials Energy Solid Waste Emissions Direct measurement ' ' Company purchasing/utility records Calculated from equipment specs Engineering estimates Permit limit Other Has extrapolation or allocation been used during data collection? If so, in what way? Figure 3.4 Data collection form. Examples of data collection sheets [2]. Key information that must be gathered and documented include: Reporting basis, time period, geographic location, and other use- ful information for characterizing the process or reporting facility, such as its age, share of total company or industry production, etc. Types and quantities of material inputs Incoming transportation of material inputs (reported by transpor- tation mode and distance) 54 LIFE CYCLE ASSESSMENT HANDBOOK • Types and quantities of useful outputs (product of interest, co-products, recycled scrap, recovered heat, etc.) • Types and quantities of water inputs (emphasis on consumptive use of fresh water) • Types and quantities of solid wastes and disposition of each type of waste (e.g., landfilled, burned, burned with energy recovery, land applied, etc.) • Types and quantities of emissions to air • Types and quantities of emissions to water • Types and quantities of materials used to package outgoing product For emissions, it is preferable to gather data on process emissions released to the environment after any on-site controls or treatment have been applied. If the reported emissions include fuel combustion emissions, it is important that this be noted, so that fuel-related emissions are not double-counted later when constructing the life cycle inventory model and linking to data sets for the reported process fuels. Emissions should be speciated to the extent possible in order to facilitate subsequent impact assessment. Instead of reporting a group of emissions as volatile organic compounds (VOC), the chemical composition of the emissions should be reported. Different isomers of a chemical can have different human health and ecotoxicity impacts (e.g., ortho- and para-xylene), so it is desirable to speciate emissions as precisely as possible. Cut-Off Criteria. Criteria for excluding components or materials are defined at the outset of the project but may change based on limitations encountered as the study is conducted. Cut-off rules are typically expressed in terms of mass, for example, "the study will account for at least 95% of the total mass of inputs, and no input shall be excluded that individually contributes 1% or more of the mass/ ' Ideally, a life cycle study would account for all life cycle steps and 100% of the content of product, modeled using data for the actual materials and pro- cesses. Practically, data are often not available for some processes or materials or cannot be gathered within the time and budget constraints for the study. This is often true in comparative analyses where the organization sponsoring the study can provide detailed data on their own system, but alternative sys- tems must be modeled using publicly available data. Before deciding to exclude materials or processes from the study, it is impor- tant to carefully consider the potential effect on study results. Mass contribu- tion is usually the criterion used to identify components for possible exclusion, but a material with a small mass contribution may have significant impacts on energy or environmental impacts. For example, an exterior metal plating a few microns thick may add only a tiny amount to the mass of a product. However, the energy and emissions associated with the production of the metal, or metal emissions from the plating process, may have impacts that are large relative to the mass of material used in the product. Another example would be a thin exterior coating that is cured in an energy- intensive baking process. Even if data on production of the specific coating LIFE CYCLE INVENTORY MODELING IN PRACTICE 55 material is not available, energy for the baking process should be included because of its contribution to energy impacts. Sometimes additional research can uncover options for inclusion of compo- nents for which data do not initially seem to be available. Take the example of a molded plastic product that has inputs of a large quantity of resin and a much smaller quantity of a material identified as a color compound. Because the color compound represents a small percentage of the total mass of inputs, and data on production of pigments are generally not available, the first inclination may be to exclude the color compound. However, manufacturer literature for the compound indicates that the compound consists of a small amount of pigment blended into the same resin used as the main input to the process. In this situa- tion, the amount of resin in the color compound can be included in the modeling, increasing the total mass of inputs covered in the analysis. Only the very small amount of pigment in the color compound would still be excluded. Information in manufacturer literature and material safety data sheets (MSDS) can often help identify constituents of specialty materials so that at least some of the content of these products can be included or suitable surrogate data can be selected. Another cut-off decision is the decision whether or not to include capital goods and infrastructure, i.e., the buildings and equipment used to manufac- ture the product, or the vehicles used to transport products. In the past, sample calculations suggested that the contributions of infrastructure tend to be small when allocated over the total amount of output over the lifetime of the infra- structure. However, more recentanalyses have indicated that infrastructure contributions may be significant for certain industry sectors, e.g., non-fossil electricity generation [3]. Constructing infrastructure models can be very com- plex and time-consuming on a study-specific basis; however, LCA databases such as ecoinvent contain general modules for commonly used infrastructure such as chemical plants, metalworking equipment, vehicles, etc. that can be used to estimate the contribution of infrastructure [4]. 3.4 Methodology Issues In this section, several important methodological issues are discussed, includ- ing areas where different methodological choices can be made depending on the specific systems being studied, as well as methodological issues that are the focus of increasing interest in the environmental community, such as water use and carbon tracking. In cases where the choice of methodology has a strong influence on the study results and conclusions, the practitioner should justify the reasons for the methodology chosen, and a sensitivity analysis should be conducted to see if an alternative methodological choice produces similar or different results and conclusions. 3.4,1 Feedstock Energy For a complete energy accounting, an LCA must include not only the energy that is expended during the process and transportation steps over the life cycle 56 LIFE CYCLE ASSESSMENT HANDBOOK of a product system, but also the energy associated with the material content of the product. Feedstock energy refers to the energy value of resources extracted from nature that are used as material feedstocks for product systems. For example, crude oil and natural gas are material feedstocks for production of traditional plastic resins, trees are harvested as feedstock for paper and lumber products, and palm oil extracted from palm fresh fruit bunches is used as a feedstock for surfactants and detergents. Franklin Associates, the original LCA firm in North America, has tradition- ally used the term "energy of material resource" (EMR) to refer to feedstock energy. Originally, EMR was used by Franklin Associates to track the energy value of fossil fuel resources that were diverted from their primary use as energy resources for use as material inputs. EMR was not tracked for biomass materials, for two reasons: (1) biomass materials were primarily used as mate- rial feedstocks or food sources rather than fuel resources, and (2) biomass is a renewable energy source, so that use of biomass for energy did not result in a net depletion of finite fuel reserves. As biomass materials are increasingly being utilized for energy purposes, the energy value of biomass materials is becoming an increasingly important issue. In addition, including feedstock energy for all types of resources, whether fossil or renewable, provides a more consistent and complete energy accounting approach. Therefore, it is good practice to track feedstock energy for all material inputs, while continuing to distinguish between renewable and non-renewable feedstocks. Different bases can be used for assigning and tracking feedstock energy. Feedstock energy is usually based on the higher or lower heating value of the material at the point of extraction from nature. The amount of energy remain- ing in the finished product is less than the total energy content of the materi- als extracted from nature, since there are losses during the steps required to convert the raw material into a finished product. In contrast to process and transportation energy that is irretrievably expended when fuel is combusted, the feedstock energy embodied in the finished product represents potentially recoverable energy. The energy remains embodied in the material as long as the material remains in use, in the original product or in recycled applications. Energy can be recovered from the material through end of life management processes, such as waste-to-energy combustion. If biomass feedstock energy is assigned to the material at the point of removal from nature, there can be additional challenges in tracking feedstock energy through subsequent processing steps. Consider the example of trees harvested as feedstock for paper production. The mass of roundwood logs brought to the mill includes the moisture content of the wood as well as the weight of bark. Bark and wood chips generated during subsequent processing of the logs are used as an energy source at the mill, and black liquor that is burned to provide energy at the mill contains lignin extracted from the wood fiber during the chemical pulping process. If feedstock energy is assigned to the total weight of wood entering the mill, then one must be careful not to double count the process energy from bark, chips, and black liquor derived from the same mass of incoming wood. LIFE CYCLE INVENTORY MODELING IN PRACTICE 57 3.4.2 Multi-Output Processes Many processes produce more than one useful output. This makes it necessary to use some method to divide or partition the process input and output flows among the useful outputs. The preferred hierarchy for handling multi-output processes is defined in ISO 14044 section 4.3.4.2 as 1. Avoid allocation where possible, 2. Where allocation cannot be avoided, allocate inputs and out- puts among useful coproducts in a way that reflects physical relationships, 3. Where physical relationship alone cannot be used, allocate based on other relationships. Two options are given for avoiding allocation. The first option is to further subdivide the given process into subprocesses with inputs and outputs that can be assigned to individual co-products. This approach can be used, for example, when operating data on a manufacturing facility are provided as a "black box," but individual co-products can be traced to separate processes within the facility. In many cases, however, even at the most detailed subpro- cess level, a single process produces multiple co-products. The second option is to avoid allocation by expanding the system boundar- ies1. In this approach, all process burdens are assigned to the primary product of interest, and credit is given for materials or services that are displaced by the other co-products of the process. System expansion may not be a suitable approach if the process being evaluated is the primary or only commercial route for producing one or more of the co-products, so that there is not a basis for displacement credit. A hydrocracker unit uses inputs of refined oil and natural gas to produce outputs of ethylene, propylene, other hydrocarbons, fuel gas, and heat. An energy credit can be applied for the fuel gas and heat co-products used outside the system boundaries, but the remaining process burdens must be allocated among the material co-products. When allocation cannot be avoided, physical relationships such as mass or energy are commonly used as the basis for allocating process flows among useful outputs. The allocation should be related to the function of the prod- ucts. Using the hydrocracker example, the ethylene, propylene, and other co-products are used as material inputs for production of plastic resins and other petrochemicals rather than as fuels, so burdens can be allocated among the co-products based on their mass rather than on their energy value. Economic value may also be used as a basis for allocation; however, this approach should be used with caution, since fluctuations in price of co-products can change the results and conclusions of an analysis, even if there has been no change in the physical relationship between the co-products. 1 System expansion is described in EPA/600/R-06/060 Life-Cycle Assessment: Principles and Practice. This approach is also referred to as the "avoided burden" approach. 58 LIFE CYCLE ASSESSMENT HANDBOOK 3.4.3 Postconsumer Recycling There are a number of approaches that can be used for modeling the impacts of postconsumer recycling.Commonly used approaches include system expan- sion, boundaries drawn between successive useful lives of the material, and allocated approaches. System Expansion. This approach avoids the need for allocation. For mate- rial that is recycled at end of life, the system that produced the material is assigned the burdens for collection of the postconsumer material and repro- cessing the recovered material into a form that is ready for its next use. Credit is then given for avoided production of the material that is displaced by the recycled material. The credit is given to the system producing the recycled material. The rationale is that because the first system is supplying material as a feedstock for future systems, less virgin material must be produced. Cut-Off Method. This approach also avoids the need for allocation. Distinct boundaries are drawn between systems producing and using recycled mate- rial. The initial system is assigned all virgin production burdens for the mate- rial, and material going to recycling leaves the first system's boundaries at end of life. The user system (second system) burdens begin with collection and reprocessing of postconsumer material.2 Since collection and reprocessing burdens are generally much lower than virgin production burdens, this approach tends to favor the system using recy- cled material. The rationale can be expressed as follows: Because the second system is using recycled material, demand for virgin material is reduced and less virgin material must be produced. Allocated Burdens. The rationale for the allocated approach is that all prod- uct systems using a given quantity of material should share equally in recy- cling burdens and benefits. This includes the system first using the material in virgin form as well as all subsequent systems using the material after recovery and reprocessing. The following equation can be used to illustrate the general concept of the allocated approach: V / n + F + U + ( n - l ) / n x R + D / n where V = virgin production, F = fabrication, U = use, R = recycling, D = dis- posal, and n = the total number of useful lives of the material (including virgin use and all subsequent uses until the material is disposed). The recycling allocation factor (n-1) is one less than the total number of uses since there is no recycling preceding the initial use. In reality, with each recycling cycle there will be collection and reprocessing losses, and subsequent uses of the recycled material are likely to be in products with a mix of virgin and recycled content. To illustrate the general concept, 2 This method is outlined in EPA/600/R-92/245 Life-Cycle Assessment: Inventory Guidelines and Principles, where it is identified as recycling allocation method 2. LIFE CYCLE INVENTORY MODELING IN PRACTICE 59 however, we will use the simplified equation, with 100% recovery and recy- cling after each use. Allocation equation applied to open-loop recycling: Open-loop recycling describes a system in which a product is recovered at the end of its useful life, and the recovered material is then used in a different type of product system. Typically the second product is disposed after use, or the mate- rial may be recovered and reused in a product that has a low recycling rate and therefore a low probability of repeated use cycles. Open-loop recycling often applies to materials with properties that degrade with repeated use cycles, for example, paper fibers that become shorter with each repulping and remanufacturing cycle. Open-loop recycling also applies to products that have low recycling rates even though the material properties may be suitable for repeated use cycles. In open-loop recycling, the total number of useful lives of the material "n" is a small number. For example, if material is used in a virgin product, recovered and recycled into a second product, and the second product is disposed at end of life, then n=2, and each system that uses the material is allocated half the virgin production burdens, recycling burdens, and disposal burdens. Allocation equation applied to closed-loop recycling: Closed-loop recy- cling occurs when material is used in a product, the product is recovered at end of life, and the recovered material goes back into the same type of product, so that there are repeated recovery and reuse cycles. In order to get a large number of use cycles out of the material, the material properties must hold up through repeated use cycles (e.g., glass, metals). As the total number of uses "n" increases, the virgin production burdens and disposal burdens allocated to each use become smaller, and (n-l)/n (the allocation factor for recycling burdens) approaches 1. There are limitations to the allocated recycling approach. This method requires assumptions about the total number of lifetime uses of the mate- rial. For a given product application, it is only possible to state with certainty whether incoming material is virgin material or postconsumer material. If the material enters the system as postconsumer material, it has had at least one previous life, but it is not possible to determine the total number of previous uses. Similarly, if the current product is known to be recycled at end of life, the material will have at least one subsequent use, but the fate of the material after the next use is uncertain. For durable products that have very long useful lives, there is additional uncertainty about future recycling. If the product is in use for many years, recycling rates and technologies at the end of the product's useful life may be quite different from recycling rates and practices at the time the product was manufactured. Allocation calculations can become very complicated when adjusting for reprocessing losses and sequential useful lives that have different mixes of 60 LIFE CYCLE ASSESSMENT HANDBOOK virgin and postconsumer content and different postconsumer recovery rates. Because the system expansion and cut-off approaches focus on the material's use in the current product system and the material's next use, the calculations for these methods are more straightforward and require fewer assumptions about prior and future uses of the material. For recycling to be sustainable, there must be a balance between recycled material supply and demand. If the supply of recycled material used by the system is currently fully utilized (e.g., secondary aluminum), then the recy- cling rate for a product must be equal to or greater than its recycled content in order to be sustainable. If a system uses more of a fully utilized recycled material than it produces (recycled content > recycling rate), then it creates a net deficit in the recycled material supply that has to be made up with virgin material. If the system's recycling rate is higher than its recycled content, then it is a net producer of recycled material, and a credit can be applied for the vir- gin material that is displaced by the surplus recovered material. 3.4.4 Converting Scrap Scrap that is generated during material converting processes is referred to as postindustrial or preconsumer scrap. Unlike postconsumer scrap, pre- consumer material has not had a previous useful life in a product, so there is no previous life to allocate virgin material burdens. However, the mate- rial is often degraded to some extent during the converting process. For example, the material may have been coated, had colorants added, or been glued or laminated to other materials. The industrial scrap material usually requires some degree of reprocessing before it can be used to produce a use- ful product. If the converting scrap is utilized internally at the same facility in the same process that produced the scrap, then this internal recycling simply reduces the net amount of virgin inputs required per unit of product output, and no allocations are needed. The process burdens for manufacturingthe primary product should include the added burdens for any reprocessing of internal scrap before it is returned to the process (e.g., regrinding of plastic molding scrap before it is put back into an extruder). If the scrap is used outside the boundaries of the system that produces the scrap, then there are different approaches that can be used to allocate the virgin material burdens associated with the scrap material. It is important to distin- guish between the burdens associated with production of the material content of the scrap and the burdens associated with the process that generates the scrap. The burdens for the converting process that generates the scrap should be assigned to the primary product, since the converting process adds no value to the scrap material and usually reduces its value. For example, the processes of applying coating to cartonboard and cutting it into carton blanks are done for the purpose of producing a finished carton blank, so the environmental burdens for these processes should be assigned to the carton blanks. The coat- ing on the trim scrap generated from the converting process reduces the value LIFE CYCLE INVENTORY MODELING IN PRACTICE 61 of the scrap, because the coated scrap must be repulped and the fiber separated from the coatings before the fiber can be used to make a useful product. When deciding whether it is appropriate to apply credits for using postin- dustrial scrap, one should consider the current utilization of this type of scrap. If the scrap is currently being fully utilized (e.g., kraft clippings from box manufacture), use of the scrap diverts the material from some other use rather than diverting it from disposal. Material production burdens can be allocated between the primary product and the scrap based on the mass of input mate- rial that ends up in the finished product and the mass of scrap that leaves the system boundaries as input to another product. However, if the scrap requires additional reprocessing, as in the example of coated cartonboard scrap, then the scrap material will be used only if the extra reprocessing steps can be economically justified. In this situation, virgin mate- rial burdens may be allocated between the primary product and the converting scrap based on the relative quantities of material that end up in the primary product and in the converting scrap, and the relative economic value of the virgin material and the devalued scrap. 3.4.5 Water Use In recent years water use, in particular freshwater use, has become an area of increased interest in life cycle inventories and assessments [7]. However, a large body of life cycle data has been developed over the years without gath- ering accompanying water use data, making it necessary to add water use to data sets where it is missing. It is interesting to note that a similar situation existed when global warm- ing became recognized as an important environmental issue. Prior to that time, carbon dioxide emissions had not been tracked in LCI databases because there had been no environmental reporting requirements for carbon dioxide. Adding carbon dioxide emissions to LCI databases was relatively easy, how- ever, since carbon dioxide emissions can be estimated based on the carbon con- tent of material. Water use, on the other hand, is more difficult to characterize and quantify, since there are many forms of water use and different types and sources of water that can be utilized. Water use is a broad term that can include any form of use that makes water permanently or temporarily unavailable for use by another system. Water use can be generally classified as in-stream use or off-stream use [7]. Off-stream use involves withdrawal from a water source, while in-stream uses do not. Examples of in-stream use are hydroelectric generation, water transport, fish- eries, or recreational uses such as boating. Off-stream uses involving water withdrawals can be further classified as degradative or consumptive. Degradative use returns the water to the same watershed from which it was withdrawn, but with changes in quality (e.g., addition of contaminants, temperature changes). Consumptive use refers to water that is withdrawn from one source and returned to a different water body or watershed (e.g., depleting the initial water source but adding to the 62 LIFE CYCLE ASSESSMENT HANDBOOK receiving water), or water that is withdrawn and not directly returned to a receiving body (such as water embodied in a product or evaporated in a cool- ing tower or drying operation). Consumptive use of water can be a very signif- icant environmental concern, particularly in areas where fresh water is scarce. There is less focus on consumptive use of saltwater. Protocols are being established for categorizing and reporting the vari- ous types of water use. There is an active UNEP/SETAC working group on water use and consumption within LCA. As of June 2011, the International Organization for Standardization initiated a working draft of a new standard, ISO 14046 Life cycle assessment - Water footprint - Requirements and guide- lines, to provide internationally harmonized metrics for water footprints [9]. 3.4.6 Carbon Tracking Considerations Biogenic carbon and fossil carbon are treated differently in life cycle meth- odology. Biogenic carbon is carbon that is removed from the atmosphere and incorporated into the physical mass of a plant or organism. The carbon remains embodied in the biomass-derived product throughout its useful life. At end of life, some or all of the carbon may be permanently sequestered (e.g., if the biomass-derived product is landfilled and some of the material does not decompose), or carbon content may be returned to the atmosphere through decomposition or combustion. Biogenic carbon released at end of life as carbon dioxide returns to the atmosphere in the same form as which it was removed, with no net increase in atmospheric carbon dioxide within the time frame of natural biogenic carbon cycling. These biomass carbon dioxide emissions are considered to be "carbon neutral." Although fossil fuels such as petroleum and coal originated as biogenic car- bon, the carbon uptake occurred millions of years ago. Fossil carbon resources remain stored within the earth until they are extracted through human inter- vention. Therefore, emissions associated with combustion of fossil fuels or fossil-derived materials are treated as net contributions to atmospheric carbon dioxide levels, and no carbon storage credit is given when materials such as fossil fuel-derived plastics are landfilled [10]. There are additional considerations regarding decomposition of landfilled biomass products. If a biomass-derived product decomposes aerobically, the carbon dioxide released is considered carbon neutral. However, if the biomass decomposes anaerobically, both carbon dioxide and methane will be produced. For either type of decomposition, the carbon dioxide produced is considered carbon neutral, but the methane is not. Since human intervention in the bio- mass carbon cycle is responsible for some of the atmospheric carbon returning to the atmosphere as methane, with a higher global warming potential than the carbon dioxide initially taken up by the biomass, the methane releases are not considered carbon neutral. End of life carbon tracking calculations can become quite complicated when considering the potential mix of fates of biomass products and the time frame over which releases occur. As noted previously, biomass-derived products may LIFE CYCLE INVENTORY MODELING IN PRACTICE 63 decompose in landfills, but this is subject to landfill conditions (e.g., tempera- ture, moisture, presence of microbes). It may take many years for the decom- position to occur, and the decomposition may never completely convert all the carbon content to carbon dioxide and methane. Samples of newspaper and other bio-derivedproducts excavated from actual landfills have shown very little degradation [11]. Landfill simulation studies have also indicated that the lignin content of products derived from woody biomass tends not to decom- pose [12]. Biomass decomposition can also be inhibited by moisture-resistant coatings, fillers and additives, or sandwiching biomass layers between layers of foil or plastic. If landfilled biomass does decompose anaerobically, there are different pos- sible fates for the methane that is generated. If the methane escapes into the atmosphere uncaptured and untreated, it results in additional global warming potential. If the methane is captured and flared (with or without energy recov- ery) or oxidizes as it travels through the landfill cover, then the carbon content returns to the atmosphere as carbon-neutral COr If the captured methane is burned with energy recovery, then the useful energy recovered can displace natural gas or electricity consumption, and credit should be given for the dis- placed energy and emissions. Because of the many uncertainties surrounding biomass decomposition in landfills, it is advisable to conduct sensitivity analyses on the carbon storage and releases associated with landfilled biomass products. Similar carbon tracking issues apply to waste-to-energy combustion of mate- rials. Carbon dioxide from combustion of biomass-derived material is consid- ered carbon neutral, while carbon dioxide from the combustion of materials derived from fossil fuels is considered as a net contribution to global warm- ing potential. Regardless of whether the carbon in the combusted material is biogenic or fossil carbon, credit should be given for the energy and emissions displaced by energy recovered from combustion of the material. 3·5 Evolution of LCA Practice and Associated Issues In recent years, LCA practice has evolved rapidly, from a specialty field practiced by a handful of practitioners with closely guarded databases, to a widely used tool with emphasis on transparency and sharing of data. Life cycle inventory data are publicly available at various levels of detail in inter- national databases, national databases, and from industry associations. A few examples of publicly available life cycle inventory databases include the European Commission ELCD database, the U.S. LCI database, and plas- tic resin databases published by European and U.S. plastics industry asso- ciations [13,14,15,16]. Although LCA practice still requires a high degree of expertise and knowledge, the availability of sophisticated LCA software such as SimaPro and GaBi have made LCA accessible to a much wider user base [17,18]. 64 LIFE CYCLE ASSESSMENT HANDBOOK The extensive databases available within LCA software tools can greatly streamline the time and effort required to conduct an LCA. However, wide- spread use of the software and public databases brings with it a new set of responsibilities and issues for consideration by the practitioner. When con- structing LCA models with these tools, it is still the responsibility of the prac- titioner to carefully review the data sets and make adjustments as necessary to ensure consistency and relevance for the systems that are being modeled. The practitioner also needs to be aware of underlying differences that may exist in different data sets that are available. Examples include: • Differences in energy content of inputs from nature. Some data- bases use the lower heating value of materials, while others use higher heating value. It is also important for the practitioner to ensure that energy flows are linked to data sets corresponding to the corresponding form or stage of material input (e.g., uranium has different energy values at different stages of processing). • Differences in methodology. Data sets may have embedded allo- cations that may not be consistent with the allocation approach used for other processes or materials in the system being mod- eled. Depending on the level of detail available in the data set, it may or may not be possible to apply a different allocation method. • Differences in naming conventions. When working within LCA software, before adding a new substance to a data set, the practi- tioner should check to make sure that the substance is not already listed in the database under a different name. Chemical Abstracts Service (CAS) registry numbers are often the most effective way to search for substances that may go by multiple names. If new flows are added, the practitioner must make sure that the added flows are also added to the appropriate impact categories and methods within the software tool. Otherwise, the flow may show up in the inventory but not in the relevant impact results. • Differences in emissions lists for the same process. For the same process, data sets from different databases or representing differ- ent countries may have variations in emissions lists. It is impor- tant to check whether these variations are associated with actual process differences or differences in completeness of reporting. The absence of an emission may not mean that emissions of that substance are zero but rather that the emission was not reported in the data sources used to develop the data set. Differences in completeness can be misleading when they result in apparent differences in impacts for systems that use similar processes and materials. • Regional differences • Differences in technology. As noted previously, technologies and technology mixes may differ in various regions of the world. For example, there are three basic electrolysis technologies for LIFE CYCLE INVENTORY MODELING IN PRACTICE 65 production of sodium hydroxide: diaphragm cell, mercury cell, and membrane cell. The percentage of European produc- tion by each technology reported in the ecoinvent database is different than the U.S. technology mix reported in the U.S. LCI database. Where possible, the mix of process technologies should be adjusted to represent the mix relevant to the system or region being modeled. Differences in material sourcing. The source or mix of feed- stock materials used for a process may be different in different geographic regions. For example, data gathered from U.S. resin producers for the U.S. LCI Database indicated a different mix of crude oil and natural gas used as material feedstocks com- pared to the feedstock mix used by European plastic producers surveyed for PlasticsEurope's resin modeling [15,16]. In addi- tion to differences in types and quantities of materials used, material sourcing differences can also affect the modeling of material transport distances and modes. Differences in electricity grids. When using process data from one region to represent the same process in a different region, it is important to link process electricity requirements to the relevant electricity grid for the region where the process is tak- ing place. 3.6 Conclusion The science of LCA continues to evolve, encompassing more environmental flows at increasing levels of detail. The LCI serves as the foundation for the subsequent impact assessment and interpretation stages of the LCA. In order for the results and conclusions of the LCA to serve as a basis for sound envi- ronmental decisions, it is essential that the LCA practitioner stay abreast of methodological developments and follow accepted LCI methodology and best practices when conducting the scoping and inventory stages. References 1. International Organization for Standardization. ISO 14040:2006(E). Environmental manage- ment - Life cycle assessment - Principles and framework. 2. International Organization for Standardization. ISO 14044:2006(E). Environmental manage- ment - Life cycle assessment - Requirements and guidelines. 3. Frischknecht, R., Althaus, H.-J., Bauer, C, Doka, G., Heck, T., Jungbluth, N., Kellenberger, D., and Nemecek, T. 2007. "The environmental relevance of capital goodsin life cycle assess- ments of products and services." Int J Life Cycle Assess. 4. Ecoinvent database. Swiss Centre for Life Cycle Inventories, http://www.ecoinvent.ch/ 5. U.S. Environmental Protection Agency (EPA). 1993. Life-Cycle Assessment: Inventory Guidelines and Principles. EPA/600/R-92/245. 66 LIFE CYCLE ASSESSMENT HANDBOOK 6. U.S. Environmental Protection Agency (EPA). 2006. Life-Cycle Assessment: Practice and Principles. EPA/600/R-06/600. 7. Koehler, Annette. 2008. "Water use in LCA: managing the planet's freshwater resources/' Int J Life Cycle Assess 13: 451^155. 8. Owens, J.W. (2002). "Water resources in life-cycle impact assessment: Considerations in choosing category indicators." Journal of Industrial Ecology 5 (2): 37-54. 9. Information from ISO website http://www.iso.org/iso/iso_catalogue/catalogue_tc/cata logue_detail.htm?csnumber=43263, accessed August 2011. 10. U.S. Environmental Protection Agency. Solid Waste Management and Greenhouse Gases: A Life- Cycle Assessment of Emissions and Sinks. 3rd Edition. September 2006. 11. Rathje, W.L., Hughes, W.W., Wilson, D.C., Tani, M.K., Archer, G.H., and Jones, T.W, The Garbage Project, Department of Anthropology /Bureau of Applied Research in Anthropology, University of Arizona, Tucson, AZ 85721 12. Eleazer, W.E., Odle, W.S. Ill, Wang, Y.S., and Barlaz, M.A. 1997. "Biodegradability of munici- pal solid waste components in laboratory-scale landfills." Env. Sei. Tech. 31(3): 911-917. 13. European Commission - Joint Research Centre - Institute for Environment and Sustainability. ELCD database, http://lct.jrc.ec.europa.eu/assessment/data 14. U.S. LCI Database, hosted by the National Renewable Energy Laboratory, www.nrel.gov/lci 15. American Chemistry Council Plastics Division resin data. Report and appendices available at http: / / plastics, americanchemis try. com/ Education-Resources/Publica tions#Resource- EnvironmentalProfileAnalyses 16. PlasticsEurope Eco-profiles. http://www.plasticseurope.org/plastics-sustainability/eco- profiles.aspx 17. SimaPro LCA Software. PRe Consultants, http://www.pre.nl/content/simapro-lca- software/ 18. GaBi LCA Software. PE International, http://www.gabi-software.com 4 Life Cycle Impact Assessment Manuele Margni1 and Mary Ann Curran2 1CIRAIG, Montreal, Canada 2US Environmental Protection Agency, Cincinnati, OH, USA* Abstract The overarching purpose of life cycle impact assessment (LCIA) is to pro- vide additional information to assess life cycle inventory (LCI) results and help users better understand the environmental significance of natural resource use and environmental releases. An important distinction exists between LCIA and other types of impact analysis, such as traditional risk assessment: LCIA does not directly assess the impact of chemical releases. This chapter presents a brief history of the development of the state-of-the-art LCIA approach, and describes the diversity that is found in the various LCIA models that are currently used, along with the accompanying criticism that comes from having several choices of methodologies. The chapter concludes with a discussion on the direction of future LCIA development. Most significantly, further development is needed in model- ing important resource-related impact categories, such as water use and land use, addressing issues such as ecosystem services, and incorporating spatial and tempo- ral differentiation. Keywords: Life cycle assessment, life cycle impact assessment, indicator, midpoint, endpoint, area of protection 4.1 Introduction A well done life cycle inventory consists of a large quantity of data about natu- ral resource use and releases to the environment. However, at this point in the assessment, these data are difficult to interpret. It is impossible to decide what the environmental impacts of a system are by considering only the mass that is extracted or released. Obviously, one pound (or kilogram) of one type of air emission or waterborne pollutant can have a vastly different impact on human * The views expressed in this chapter are those of the authors and do not necessarily reflect the views or policies of the US Environmental Protection Agency. Mary Ann Curran (ed.) Life Cycle Assessment Handbook: A Guide for Environmentally Sustainable Products, (67-104) © 2012 Scrivener Publishing LLC 67 68 LIFE CYCLE ASSESSMENT HANDBOOK health and the environment than another. Also, with varying sites where the release occurs, an amount of one pollutant can have different effects under dif- ferent condition. The life cycle impact assessment (LCIA) phase of an LCA is the evaluation of potential human health and environmental impacts of the natural resources and environmental releases identified during the inventory. By modeling poten- tial impact pathways, LCIA addresses ecological and human health effects, as well as resource depletion, in order to help us better understand the linkage between the product or process and its potential environmental impacts. An important distinction exists between LCIA and other types of impact analysis (see box). LCIA does not necessarily attempt to quantify site-specific or actual impacts associated with a product, process, or activity. That is, LCIA does not attempt to directly assess the impact of releases as a traditional risk assess- ment would. According to the US Environmental Protection Agency (EPA), "Risk assessment is a process in which information is analyzed to deter- mine if an environmental hazard might cause harm to exposed persons and ecosystems (EPA 2004)." Box 4.1 LCA versus Risk Assessment (EPA 2006) "An important distinction exists between life cycle impact assessment (LCIA) and other types of impact analysis. LCIA does not necessarily attempt to quantify any specific actual impacts associated with a product, process, or activity. Instead, it seeks to establish a linkage between a system and potential impacts. The models used within LCIA are often derived and simplified versions of more sophisticated models within each of the various impact categories. These simplified models are suitable for relative compar- isons of the potential to cause human or environmental damage, but are not indicators of absolute risk or actual damage to human health or the envi- ronment. For example, risk assessments are often very narrowly focused on a single chemical at a very specific location. In the case of a traditional risk assessment, it is possible to conduct very detailed modeling of the predicted impacts of the chemical on the population exposed and even to predict the probability of the population being impacted by the emission. In the case of LCIA, hundreds of chemical emissions (and resource Stressors) which are occurring at various locations are evaluated for their potential impacts in multiple impact categories. The sheer number of Stressors being evaluated, the variety of locations, and the diversity of impact categories makes it impossible to conduct the assessment at the same level of rigor as a traditional risk assessment. Instead, LCIA models are based on the accepted models within each of the impact categories using assumptions and default values as necessary. The resulting models that are used within LCIA are suitable for relative comparisons, but not sufficient for absolute predictions of risk." LIFE CYCLE IMPACT ASSESSMENT 69 Instead of trying to connect direct hazard and harm, LCIA seeks to establish a linkage between a system and potential impacts. The key concept in this component is that of Stressors - a set of conditions that may lead to one or multiple impacts. 4.2 Life Cycle Impact Assessment According to ISO 14040-44 Requirements 4.2.1 Overview According to the International Organization for Standardization1 (ISO) 14040 and 14044 standards on LCA, LCA addresses the environmental aspects and potential environmental impacts (e.g. depletion ofresources and the environmental conse- quences of releases) throughout a product's life cycle from raw material acquisi- tion through production, use, end-of-life treatment, recycling and final disposal (i.e. cradle-to-grave). LCIA is the third out of four inter-related phases in an LCA study (the other phases are Goal & Scope (G&S) definition, Life Cycle Inventory (LCI) analysis, and Interpretation). While this latter step is transversal to the other three, the LCIA is subsequent to the first two steps (see Figure 4.1). Life cycle assessment framework Goal and scope definition . Al J u Inventory analysis J u Λ Impact accpccrnpnt ^ W\ Interpretation Direct applications: - Product development and improvement - Strategic planning - Public policy making - Marketing - Other Figure 4.1 Life Cycle Assessment (LCA) is comprised of four inter-related phases (ISO 2006a; ISO 2006b). ISO is also referred to as the International Standards Organization. 70 LIFE CYCLE ASSESSMENT HANDBOOK The overarching purpose of LCIA is to provide additional information to help assess a product system's LCI results to better understand their environ- mental significance. In other words the LCIA step is intended to be a way to evaluate the significance of the environmental interventions of an LCI and support its interpretation within the given project scope. It is, therefore, not the primary purpose of LCIA to calculate an absolute value of an environmental indicator or a set of environmental indicators, but it is to determine the relative importance of each elementary flow within a given environmental problem, i.e. one of multiple impact categories, and aggregate them into a manageable set of indicators. When interpreting the LCIA results, it is of primary impor- tance to keep in mind that the absolute values of these LCIA indicators do not predict absolute or precise environmental impacts due to: • The relative expression of potential environmental impacts to a reference unit, • The integration of environmental data over space and time, • The inherent uncertainty in modeling environmental impacts, and • The fact that some possible environmental impacts may occur in the future. The LCIA phase could be compared to converting currency when consoli- dating the accounting of an international company. In this analogy, the G&S defines what should be included in the accounting plan, and the LCI step con- sists of accounting for all incomes and expenses and reporting them in dif- ferent monetary units. Thus, the LCIA step is the conversion of the different national currencies into a single currency to be used to consolidate the accounts. Currency conversion factors are simple deterministic values, but are computed by sophisticated economic models. Similarly, characterization factors (CFs) are also deterministic numbers used as multipliers translating inventory flows into impact scores with common units representing an environmental issue. However, these numbers are often backed by calculations that use sophisti- cated natural science based models to reflect environmental mechanisms along a cause-effect chain starting from the environmental emission to an impact. According to ISO 14044 the LCIA consists of 3 mandatory elements and three optional elements (see Figure 4.2). These elements are described below. 4.2.2 Mandatory Elements Selection of Impact Categories, Category Indicators and Characterization Models. The first step within the framework of an impact analysis is the selection of impact categories in connection with defining the goal and scope of the study. The impact assessment categories should link the potential impacts and effects to the entities that we aim to protect. The commonly-accepted areas of protection (AoP) are: LIFE CYCLE IMPACT ASSESSMENT 71 LIFE CYCLE IMPACT ASSESSMENT Mandatory elements C Selection of impact categories, category indicators and characterization models Assignment of LCI results (classification) [ Calculation of category indicator results (characterization) Category indicator results, LCIA results (LCIA profile) ^^>L Optional elements Calculation of the magnitude of category indicator results relative to reference information (normalization) Grouping Weighting Figure 4.2 Elements of the LCIA phase are mandatory or optional, depending on the goal of the study (ISO 2006b). • Natural Resources • Natural Environment • Human Health • and often also, Man-Made Environment Multiple impact pathways originating from the LCI link emissions and extractions to impact category indicators. In practice, a category indicator is the outcome of a simplified model of a very complex reality, giving only an approximation of the quality status of the affected entity. Impact categories and corresponding indicators can be organized at two levels along the cause- effect chain: at a midpoint and at an endpoint level (Jolliet, Müller-Wenk et al. 2004) and (Bare and Gloria 2006). Figure 4.3 provides an example of a graphical representation of the midpoint-endpoint framework as proposed by the ILCD Handbook (EC-JRC 2010a). Assignment of LCI Results to the Selected Impact Categories (Classification). In this step, the inventory data are assigned to categories according to the impact to which they are known to relate. If a substance contributes to more than one impact category, then it is assigned to all of these categories in its entirety (that is, it is not partitioned or allocated in any way). Such a case is, for example, the 72 LIFE CYCLE ASSESSMENT HANDBOOK Midpoint Area of protection -> Human health LCI results -► Natural environment Climate change — Ozone depletion Human toxicity Respiratory inorganics Ionising radiation Noise Accidents Photochemical ozone Formation Acidification Eutrophication Ecotoxicity Land use Resource depletion Dessication salination Figure 4.3 Relationship between midpoint impact categories and three areas of protection (adapted from EC-JRC 2010a). release of nitrogen oxides that can lead to both photochemical ozone creation and acidification. The entire quantity of the nitrogen oxides would be assigned to both impact categories (i.e. 100 percent to photochemical ozone creation and 100 percent to acidification). The following example inventory data can be mapped to several impact indicators, as shown: -► Natural resources LCI Impact Category Carbon dioxide Methane< CFCs- Global Warming Potential ► Stratospheric Ozone Depletion Potential Photochemical Ozone Creation Potential Halons Nitrogen oxides Sulphur dioxide - ^ Acidification Potential Calculation of Category Indicator Results (Characterization). Classification is followed by the characterization step in which every substance is run through a model to calculate its potential impact in the impact category (or categories) to which it was assigned. The potential impact of a substance is given relative to a dominant factor in the category. For example, Climate Change potential is typically based on 1 kg of C0 2 emissions (and reported in units of C02- equivalents). These relative impacts (the CFs of a substance) are than multi- plied with the amount of each emission and the resulting impact values are summed for the respective impact category. 4.2.3 Optional Elements Depending on the goal and scope of the LCA, the following optional elements may also be implemented. LIFE CYCLE IMPACT ASSESSMENT 73 1. Normalization - calculation of the magnitude of category indica- tor results relative to reference information; 2. Grouping - sorting with the aim of possibly reducing the number of impact categories, as well as possibly ranking them in order of importance; 3. Weighting - converting and possibly aggregating indicator results across impact categories using numerical factors based on value- choices; data prior to weighting should remainavailable; and 4. Data Quality Analysis - developing a better understanding of the reliability of the indicator results in the LCIA profile. The optional LCIA elements may use information from outside the LCIA framework. The use of such information should be explained and the explana- tion should be reported. The application and use of normalization, grouping and weighting methods shall be consistent with the goal and scope of the LCA and fully transparent. All methods and calculations used shall be documented to provide transparency (EC-JRC 2010b). 4.2.4 Interpreting an LCIA Profile The interpretation phase of LCA entails the evaluation of the results of the inventory analysis along with the results of the impact assessment to aid in the decision making process, whether it is to select the preferred product, improve a process or service, etc. with a clear understanding of the uncertainty and the assumptions used to generate the results. Very seldom will the results of an LCA identify a clear "winner" between alternatives. In some cases, it may not be possible to state that one alternative is better than the others because of the uncertainty in the final results. This does not imply that efforts have been wasted or that LCA is not a viable tool for decision makers. The LCA pro- cess will still improve understanding of the environmental and health impacts associated with each alternative, where they occur (locally, regionally, or glob- ally), and the relative magnitude of each type of impact in comparison to each of the proposed alternatives included in the study. This information more fully reveals the pros and cons of each alternative. While conducting the LCA (within both the LCI and LCIA) it is necessary to apply various modeling assumptions and engineering estimates. At times these choices are based on the values held by the modeler, or by the person who commissioned the study. Therefore, every choice must be stated and the impact on the decision clearly communicated within the final results to com- prehensively explain conclusions drawn from the data. ISO (2006a) defines two objectives of life cycle interpretation: 1. Analyze results, reach conclusions, explain limitations, and pro- vide recommendations based on the findings of the preceding phases of the LCA, and to report the results of the life cycle inter- pretation in a transparent manner. 74 LIFE CYCLE ASSESSMENT HANDBOOK 2. Provide a readily understandable, complete, and consistent pre- sentation of the results of an LCA study, in accordance with the goal and scope of the study. It is important to remember that LCA is best used as an iterative approach. It is especially important to determine that if the results of the impact assessment or the underlying inventory data are incomplete or unacceptable for drawing conclusions and making recommendations, then the previous steps must be repeated until the results can support the original goals of the study. Also, LCA as a decision support tool should be used in conjunction with other decision criteria, such as cost and performance, to make a well-balanced decision. 4.3 Principles and Framework of LCIA According to ISO, the assessment of the magnitude of potential impacts on the environment is called characterization. The CF is applied to convert the results of a life cycle inventory assigned to a given impact category to the common unit of that category indicator. It is a numerical value expressing the relationship between an environmental intervention (e.g. 1 mg of lead emitted into air) and an environmental indicator. This latter is generally calculated by a characteriza- tion model that expresses a simplified mathematical representation of physical, chemical and biological processes occurring along the cause-effect chain. The collection of individual characterization models or methods (each addressing their separate impact category) is referred to as an "LCIA method- ology" (e.g. referring to the CML 2002 method or the IMPACT 2002+ method). "Method" refers, therefore, to the individual characterisation model while "methodology" is the collection of methods. According to ISO 14044, the indicator of an impact category can be cho- sen anywhere along the stressor-impact chain (i.e. the impact pathway) which links inventory data to impacts which are directly related to an AoP, i.e. Human Health, Natural Environment, and Natural Resources. Characterization can be, and is, conducted by some practitioners at the endpoint. However, more com- monly models apply CFs at the midpoint level to reflect impact calculations somewhere along (but before the end of) the impact pathway (Jolliet, Miiller- Wenk et al. 2004). A trade-off between midpoint and endpoint modeling exists. On one hand, midpoint indicators are removed from observable or tangible impacts, making it harder for the public to relate to the indicator results. On the other hand modeling to an endpoint introduces additional uncertainty as the location specific data become less certain and less available. That is, it is easier for people to grasp the significance of crop loss due to acid rain rather than an indicator that shows a potential increase in acidification, measured in hydrogen-ion equivalents. Midpoints are defined where a common mechanism for a variety of sub- stances within that specific impact category exists. Impacts that occur at the global level, such as global warming and ozone depletion, are more amenable LIFE CYCLE IMPACT ASSESSMENT 75 to midpoint modeling. Other more heterogeneous impact categories, such as "Human Toxicity" and "Ecotoxicity Effects," do not fit well into a single mid- point and, in effect, approach the AoPs. The Life Cycle Initiative, a joint project between the United Nations Environment Programme (UNEP) and the Society for Environmental Toxicology and Chemistry (SETAC) propose a comprehensive LCA frame- work that combines midpoint-oriented and damage-oriented approaches in a common and consistent framework (Jolliet et ah, 2004). While elaborating the guidance of recommended practices in LCIA, (Margni, Gloria et ah 2008) pro- vide a detailed discussion on the choice of midpoint and damage indicators. Jolliet et al. (2004) and Bare and Gloria (2006) propose dividing impact cat- egories into two groups: (1) Those based on common impact mechanisms and (2) Those that may not have a common midpoint and are comprised of different environmental mechanisms. The first group includes relatively well-established midpoints (global warming and ozone depletion) based on common impact mechanisms and for which further modeling does not differentiate between various substances. This type of impact categories is illustrated in Figure 4.4. Examples of the first type of traditional midpoint categories include: ozone deple- tion, global warming, acidification, eutrophication, and smog formation. [ DAMAGE indicator may aggregate all calculated endpoint effects into single unit. Endpoint effects not calculated or which do not have damage indicators are lost Figure 4.4 Impact models progresses from inventory flow to damage for classic midpoint impact categories. Note that endpoints that are not modeled are lost (Bare and Gloria 2006). 76 LIFE CYCLE ASSESSMENT HANDBOOK In practice many methods do not report the endpoint level (case of skin cancer), which is an interim result but reported in a damage units. These damage models can have units of Disability Adjusted Life Years (DALYs), an aggregation of environmental impacts, monetary value, or other aggregated damage units. Even though the remainder of the environmental mechanism from mid- point to endpoint to damages describes the link to environmentally relevant endpoint indicators, this sometimes occurs at the expenses of the compre- hensive nature of the midpoint, and likely resulting in higher uncertainty. In certain categories, providing methodologicalapproaches that character- ize the environmental mechanism closer to endpoints and damages does not provide additional distinction of differences in impact between substances. However, a model between damage and midpoint may add relevance (either in a quantitative or qualitative manner - in cases where quantification of endpoints is difficult to impossible), and this relevance may be added for all substances in the same way This could also enable us to compare the outcomes of different midpoint categories using models based on natural science instead of weighting factors based on social science. In a midpoint model it seems wise to minimize the unnecessary uncertainty by choosing a midpoint indicator as early as possible in the environmental chain where all substances are unified in an indicator yet the five criteria are still satisfied: comprehensiveness, relevance/reproducibility, transparency, validity and compatibility (see Section 1). The second group of impact categories, illustrated in Figure 4.5, may not have a common midpoint and are comprised of different environmental mech- anisms. Examples of the second type of impact categories which are almost always represented at an aggregated level (either at damage or midpoint level) include human toxicity and ecotoxicity, where interim human health endpoints that may be aggregated include neurological, reproductive, respiratory, and cardiovascular health endpoints. The aggregation may be in units of DALYs, monetary value, or a unitless score which is based on the relative human toxic- ity potency after including the fate, transport, and toxicity of the substances and comparing to a reference substance. The ILCD Handbook suggests considering the following points (EC-JRC 2010a): 1. For the first group of impact categories described above, the goal of damage modeling is to make results in different midpoint cat- egories comparable, and sometimes to arrive to a single score, or smaller number of environmental scores. It can then replace or support weighting practices in the midpoint approaches. The choice to stay at the midpoint level or go to the damage level is left to the user. 2. When the decision has been made to go to the damage level on an impact category of the first type (e.g., global climate change), care must be taken to ensure comprehensiveness. For example, while LIFE CYCLE IMPACT ASSESSMENT 77 DAMAGE indicator may aggregate all calculated endpoint effects into single unit. Endpoint effects not calculated or which do not have damage indicators are lost ^ y Figure 4.5 Progression from inventory flows to damage for human health. Note that endpoints not included in the damage indicators are lost (Bare and Gloria 2006). it may be relatively easy to quantify some impacts (e.g., malaria), other impacts (e.g., the impact on biodiversity) may not be so eas- ily quantified and thus may be lost. 3. Intermediary steps should be made explicit and reported sepa- rately. For example, if number of cases, Years of Life Lost (YLL) and Years of Life Disabled (YLD) are utilized then these should be considered first separately for impacts on human health. Disability weighting could then be explicitly considered if desired to group diseases together to arrive to DALY. 4. All modeling (midpoint and damage) should be properly docu- mented on data and modeling uncertainty and reliability. Value choices should be made explicit and properly documented (implicit and explicit in midpoint and damage modeling). As a matter of fact, it is important to be more specific about these values choices to decrease the uncertainty. There is no unique uni- versal set of values. In the end, LCIA approaches are typically viewed along one of two families: classical methods that determine impact category indicators at an intermedi- ate position of the various impact pathways (e.g. ozone depletion potential) or damage-oriented methods that aim to present results in the form of damage indicators at the level of an ultimate societal concern (e.g. harm to human health). 78 LIFE CYCLE ASSESSMENT HANDBOOK 4.4 Historical Developments and Overview of LCIA Methodologies The first impact assessment methodologies for LC A, termed Life Cycle Impact Assessment methodologies, can be traced back to before 1990 with the publica- tion of the Critical Volumina approach (BUS 1984). Its basic principle relied on the calculation of an equivalent volume of air, soil or water required to dilute a pollutant emission in the respective environmental media up to a threshold value, traditionally set in the regulation. Since then, significant advances have been made on: (1) comprehensively describing and modeling cause-effect chain relationships linking emissions and resources consumption to potential dam- ages; (2) improving the relevance of modeled impact pathways, (3) improving the scientific robustness of the characterization models and, (4) last but not least increasing the coverage of characterized elementary flows. For example, the first LCIA methods addressed only a few impact categories characterizing up to few dozen elementary flows. Nowadays, recent method- ologies, such as ReCiPe or IMPACT World+, are able to model up to 30 mid- point impact categories and offer the opportunity to link them to three main AoPs, providing overall thousands CFs. Figure 4.6 provides an overview of the historical development of the most common LCIA methodologies. ReCiPe (2009) was released as an update to Eco-indicator 99 and CML 2001. Similarly, IMPACT 2002+, LUCAS and EDIP will no longer be updated as the methodological developments will go toward IMPACT World+ (2012). It 1984 Volumes critiques (Bus) 1997 Ecofactors (BUWAL) 1991 Ecoscarcity (BUWAL) 1995 Eco-indicator 95 (PRe) Eco-indicator S (PR«) 2003 Impact 2002+ (EPFL) 2007 Ecoscarcity 2006 (ESU-Services, E2, FOEN, & ÖBU) 2005 LUCAS (CIRAIG) 2010 I LCD handbook (EC-JRC) 1992 (CML) EPS 2000 (CPM) 1993 EPS (CPM) 1997 EDIP 97 (IPU) 2001 (CML) 2003 TRACI (EPA) 2003 JEPIX (Japan) LIME (METI) 2004 EDIP 2003 (DTU) 2012 IMPACT world+ (CIRAIG, UM DTU, Quantis) 2009 ReCiPe (RIVM, CML, PRe, CE delft) BUS: Bundesamt für Umweltschutz BUWAL: Bundesamt für Umwelt, Wald und Landschaft CIRAIG: Interuniversity Research Center for Life Cycle of Products, Processes and Services CML: Centruum voor Milieukunde Leiden CPM: Centrum för Produktrelaterad Miljöanalys EDIP: Environment Design of Industrial Products EPFL: Ecole Polytechnique Federale de Lausanne EPS: Environmental Priority Strategies EC-JRC: European Commission-Joint Research Center FOEN: Swiss Federal Office for the Environment IPU: Instituttet for Produktudvikling JEPIX: Japan Environmental Policy Priorities Index LIME: LCIA Method based on Endpoint Modeling METI: Ministry of Economy, Trade and Industry ÖBU: Schweizerische Vereiningung für ökologisch Bewusste Unternehmensführung ReCiPe: An acronym of "RIVM, Radboud University, CML, and PRe" RIVM: Rijksinstituut voor Volksgezondheid en Milieu TRACI: Tool for the Reduction and Assessment of Chemical and other environmental Impacts UM: University of Michigan Figure 4.6 Timeline of the introduction of the most common life cycle impact assessment (LCIA) methodologies. LIFE CYCLE IMPACT ASSESSMENT 79 is interesting to note that the lifetime of LCIA methodologies (time elapsed between two releases) has varied between 5 to 10 years. On one hand, this reflects the high effervescence of research in LCIA over the last two decades, which led to rapid methodological development. On the other hand, it also reflects that LCIA is still in its infancy. From a practical point of view, any given LCIA methodology older than 10 years is no longer likely to reflect the state of the art, thereby, showing seri- ous methodological weaknesses. LCA practitioners should avoidusing older methods without at least checking the robustness of the results and conclu- sions by performing a sensitivity analysis with a more recent methodology. In the early 1990's, three methodologies were published that formed the basis of three main schools of thought that influenced the subsequent develop- ment of LCIA: • Damage-Oriented (Area of Protection): The EPS (Environmental Priority Strategies) methodology is based on a damage oriented modeling approach and expresses results in monetary values. First published in 1993 an updated version has been released in 2000 (Steen 1999). Category indicators are chosen to represent actual environmental impacts on five safeguard subjects: human health, ecosystem production capacity, biodiversity, abiotic resources, and recreational and cultural values. Weighting factors for the category indicators are determined according to people's willingness to pay and expresses the price the society is ready to pay in order to avoid these damages. • Distance-to-Target: The Swiss Ecoscarcity (or Ecopoints) approach is based on the distance-to-target principle. A first version was published in 1991and has successively been updated to 1997 (Brand, Braunschweig et ah 1997). Eco-factors were originally developed for Switzerland using the latest available statisti- cal data and the supported goals of Swiss environmental policy which set critical flows. • Midpoint-Oriented: In 1992, the Centre of Environmental Science (CML) at Leiden University produced a Guide and Background document on the LCA methodology, known as the CML 1992 methodology (Heijungs, Guinee et al. 1992). This was the first mid- point-oriented LCI methodology. Updated in 2002, CML claimed to provide best practices for operationalizing the ISO14040 series of Standards (Guinee, Gorree et al. 2002). The knowledge gained in developing these three methodologies formed the basis of LCIA methodology as it evolved. Following are brief descriptions of the main LCIA methodologies that are currently used (Bare and Gloria 2006). Additional information on these approaches can be found in the ILCD background document that analyzes existing environmental impact assess- ment methodologies for use in LCA (EC-JRC 2010). 80 LIFE CYCLE ASSESSMENT HANDBOOK CML 2002 http://www.leidenuniv.nl/cml/ssp/projects/lca2/lca2.html Developed by Leiden University's Institute of Environmental Sciences (CML) in the Netherlands, the developers of CML 2002 aimed to operationalize the ISO14040 standards and provide best practice for midpoint indicators. CML 2002 includes nine "baseline" impact categories that are used in almost all LCA studies, and twelve "study-specific impact categories" that may merit inclu- sion, if appropriate to the goal and scope. CML 2002 includes recommended methods for normalization but no recommended methods for weighting. After conducting an extensive review of existing methodologies, the authors of CML's LCA Handbook provide CFs for more than 1500 different LCI results. For most impact categories, a baseline and a number of alternative character- ization methods are recommended. In addition, a comprehensive list of CFs and normalization factors are supplied. CML 2002 Reference: Guinee, J.B. (Ed.), M. Gorree, R. Heijungs, G. Huppes, R. Kleijn, A. de Koning, L. van Oers, A. Wegener Sleeswijk, S.Suh, H.A. Udo de Haes, JA. de Bruijn, R. van Duin and MA.J. Huijbregts (2002). Handbook on Life Cycle Assessment: Operational Guide to the ISO Standards. Series: Eco-efficiency in industry and science. Kluwer Academic Publishers, Dordrecht. Eco-Indicator 99 http: / / www.pre.nl/eco-indicator99/ Developed by PRe Consultants in the Netherlands, Eco-indicator 99 is a damage-oriented approach that characterizes elementary flows into eleven midpoint categories as an intermediary modeling step toward damage modeling of three endpoint categories: human health, ecosystem quality Table 4.1 Midpoint impact categories modeled in CML 2002. Baseline: Depletion of Abiotic Resources Impacts of Land Use - Land Competition Climate Change Stratospheric Ozone Depletion Human Toxicity Ecotoxicity - Freshwater Aquatic - Marine Aquatic - Terrestrial Photo-Oxidant Formation Acidification Eutrophication Study-Specific: Impacts of Land Use - Loss of life support function - Loss of biodiversity Ecotoxicity - Freshwater sediment - Marine sediment Malodorous Air Noise Waste Heat Casualties Other: Depletion of Biotic Resources Dessication Malodorous Water LIFE CYCLE IMPACT ASSESSMENT 81 and resource depletion. The system of cultural theory separates the dam- age models into three personal perspective categories: egalitarian, long time perspective whereby a minimum of scientific proof justifies inclusion; indi- vidualist, short time perspective whereby only proven effects are included; hierarchist, balanced time perspective whereby consensus among scien- tists determines inclusion of effects. The hierarchist version is chosen as the default, while the other two versions are suggested for use in a robustness analysis. Normalization factors represent the environmental load of one aver- age European (calculated by dividing the total environmental load in Europe by the number of inhabitants and multiplying by a scaling factor of 1000). The model assumes all emissions and land uses, and subsequent damage, occur in Western Europe, except for damages to resources and those leading to climate change, ozone layer depletion, air emissions of persistent carcinogenic sub- stances, inorganic air pollutants that have long-range dispersion, and some radioactive substances. A subsequent weighting step might be performed to view results in a single score applying weighting factors specific to each cul- tural perspective (Table 4.2). Eco-Indicator 99 Reference: PRe (2001) The Eco-Indicator 99: A Damage Oriented Method for Life Cycle Impact Assessment, Methodology report. June 22, 2001. 3rd edition. PRe Consultants, Amersfoort, The Netherlands. EDIP (1997-2003) http://www.ipl.dtu.dk/English.aspx http://ipt.dtu.dk/~mic/EDIP2003 The Environmental Design of Industrial Products (EDIP) program is the result of collaborative efforts of five major Danish companies, two institutes of the Denmark Technical University (DTU), and the Confederation of Danish Industries. An update of the EDIP97 method, EDIP2003 supports spatially dif- ferentiated characterization modeling, encompassing a larger portion (through additional sub-categories of the environmental mechanisms) than EDIP97, resulting in a method providing midpoint indicators closer to a damage- oriented approach. In this respect, EDIP2003 is primarily a midpoint method Table 4.2 Weighting sets for hierarchist, egalitarian, and individualist perspectives based on panel survey. Human Health Ecosystem Quality Resources Hierarchist 40% 40% 20% Egalitarian 30% 50% 20% Individualist 55% 25% 20% 82 LIFE CYCLE ASSESSMENT HANDBOOK Table 4.3 Midpoint impact categories modeled in EDIP2003. Global Warming Ozone Depletion Terrestrial Eutrophication Aquatic Eutrophication Photochemical Ozone Formation Acidification Human Toxicity Ecotoxicity Noise with normalization (but not weighting) and considers the characteristics of the receiving environment in an effort to increase the relevance of the calculated impacts. In EDIP97, a uniform environment is assumed and is based solely on the knowledge of the emitted substance. In contrast, EDIP2003 incorpo- rates characteristics of the receiving environment in an effort to increase the relevance of the calculated impact. EDIP2003 Reference: Hauschild, M. and J. Potting (2005). Spatial differentiation in Life Cycle Impact Assessment : The EDIP03 Methodology. Environmental News No. 80. Guidelines from the Danish Environmental Protection Agency, Copenhagen, Denmark. EPS 2000 http: // cpmdatabase.cpm.chalmers.se/ AboutDatabase_2.htm The Environmental Priority Strategies (EPS) is a design tool intended to aug- ment a company's internal product development process, specifically to aid in choosing between two product concepts. Category indicators are chosen based on their suitability for assigning values to product design choices. In the EPS 2000 method, impact categories and category indicators are chosen to repre- sent actual environmental impacts on five safeguard subjects: human health, ecosystem production capacity, biodiversity, abiotic resources, and recreational and cultural values. The CF is the sum of a number of pathway-specific CFs describing the average change in category indicator units per unit of an emis- sion (e.g., kg decrease of fish growth per kg emitted S02). An estimate is made of the standard deviation in the CFs due to real variations depending on exog- enous and endogenous factors (e.g., emission location and model uncertainty). Therefore, CFs are available only where there are known and likely effects. CFs are given for emissions defined by their location, size, and temporal occur- rence. The majority of factors is for global conditions that occurred in 1990 and represents average emission rates. This means that many toxic substances, which are present mostly in trace amounts within that time frame, have a low average impact. Weighting factors for the category indicators are determined according to an individual's willingness to pay to avoid one category indicator unit of change in the safeguard subjects. Normalisation is not applied as this is a monetization approach. LIFE CYCLE IMPACT ASSESSMENT 83 Table 4.4 Midpoint impact categories modeled in EPS 2000. Human Health: Life Expectancy Severe Morbidity and Suffering Morbidity Severe Nuisance Nuisance Natural Environment: Crop Production Capacity Wood Production Capacity Fish and Meat Production Capacity Base Cation Capacity Production Capacity for Water - Irrigation Water - Drinking Water Depletion of Reserves: - Element - Fossil Oil - Fossil Coal - Mineral Extinction of Species Cultural and recreation value indicators are defined as needed. Table 4.5 Midpoint impact categories modeled in IMPACT 2002+ Global Warming Human Toxicity Respiratory Effects Ionizing Radiation Ozone Layer Depletion Photochemical Oxidation Aquatic Ecotoxicity Terrestrial Ecotoxicity Aquatic Acidification Aquatic Eutrophication Terrestrial Acidification/ Nutrification Land Occupation Non-Renewable Energy Mineral Extraction Reference for EPS 2000: CPM (1999) A Systematic Approach to Environmental Priority Strategies in Product Development (EPS). Version 2000 - General System Characteristics. CPM report 1999: 4, prepared by B. Steen, Chalmers University of Technology Gothenburg, Sweden. IMPACT 2002+ http: / / www.impactmodeling.org/ The IMPact Assessment of Chemical Toxicants (IMPACT) 2002+ methodol- ogy presents a combined midpoint/damage approach, linking all types of life cycle inventory results (elementary flows and other interventions) via fourteen midpoint categories to four damage categories: Human Health, Ecosystem Quality, Resources and Climate Change. This latter has been con- sidered representative for the Area of Protection - Life Supporting Function. For IMPACT 2002+ new concepts and methods have been developed, espe- cially for the comparative assessment of human toxicity and eco-toxicity. Human Damage Factors are calculated for carcinogens and non-carcinogens, employing intake fractions, best estimates of dose-response slope factors, as well as severities. Both human toxicity and ecotoxicity effect factors are based on mean responses rather than on conservative assumptions. Other mid- point categories were adapted from existing characterizing methods such as Eco-indicator 99 and CML 2002. The IMPACT 2002+ method presently provides 84 LIFE CYCLE ASSESSMENT HANDBOOK CFs for almost 1500 LCI results. Normalisation factors represent the environ- mental load of one average European and can be carried out either at midpoint or at damage level. References for IMPACT 2002+: Humbert S, Margni M, Jolliet O (2005) IMPACT 2002+: User Guide - Version 2.1, Swiss Federal Institute of Technology (EPFL), Lausanne, Switzerland. Jolliet O, Margni M, Charles R, Humbert S, Payet J, Rebitzer G, Rosenbaum R (2003): IMPACT 2002+: ANew Life Cycle Impact Assessment Methodology IntJLCA8(6)324-330. IMPACT World+ IMPACT World+ is an update to IMPACT 2002+. It is being developed by a consortium of researchers including CIRAIG at the Ecole Polytechnique de Montreal, Denmark Technical University (DTU), Quantis International, Ecole Polytechnique de Lausanne (EPFL), and the University of Michigan. The developers of IMPACT World+ recognize the need to offer a regionalized methodology at the global scale, implementing state-of-the-art characteriza- tion modeling developed since the publication of IMPACT 2002+ and LUCAS, and include uncertainty information encompassing both spatial variability and model uncertainty. This not only allows applying more environmentally relevant CFs, but also a regional assessment of any geo-referenced emission. This helps to ultimately determine the uncertainty related to an unknown loca- tion of an emission by associating the corresponding geographical variability to each CF at a given geographical scale. LIME http://www.jemai.or.jp/lcaforum/index.cfm The LCA National Project of Japan developed a damage-oriented (endpoint) impact assessment method called LIME (Life-cycle Impact assessment Method) that quantifies environmental impacts as a result of environmental loadings in Japan. LIME covers the potential damage on socioeconomic impacts caused by the utilization of abiotic resources, and increased extinction risk and loss of primary production caused by mining of resources measured as main dam- ages of resource consumption. Modeling socioeconomic impacts is based on the concept of user-cost which accounts for the equity of future generations. The procedure to measure damage to ecosystems is based on studies estimat- ing the risk of specific species extinction. Damage factors of mineral resources, fossil fuels and biotic resources enables LIME users to compare and integrate the damages derived from the other impact categories without the use of value judgment. For characterization, LIME involves eleven midpoint impact cat- egories. The damage assessment categories were catalogued into four areas of protection (safeguard subjects): human health, social welfare, biodiversity, and plant production. The weighting method is based on a combined analysis to provide weighting across the four areas of protection. With this analysis, two types of weighting factors were collectively implemented: (1) An amount of LIFE CYCLE IMPACT ASSESSMENT 85 monetary value for avoiding a unit amount of damage to a safeguard subject, and (2) a relative weighting coefficient based on an annual amount of damage to a safeguard subject. Reference for LIME: Itsubo N and A Inaba (2004) "LIME - A Comprehensive Japanese LCIA Methodology Based on Endpoint Modeling," in Proc. 6th International Conference on EcoBalance. ReCiPe http: / / www.lcia-recipe.net/ ReCiPe was created under a joint effort of the RIVM (Rijksinstituut voor Volksgezonheid en Milieu), CML, PRe Consultants, Radboud Universiteit Nijmegen and CE Delft. The ReCiPe approach combines the midpoint approach of Dutch CML with the damage approach of Eco-indicator 99, allow- ing users to choose which level, midpoint or endpoint, is desired for reporting indicators. The ReCiPe developers describe this as allowing the 'the user to choose between uncertainty in the indicators, and uncertainty on the correct interpretation of indicators." That is, the user can choose between eighteen rel- atively robust, but not easy to interpret, midpointsversus three easy to under- stand, but more uncertain, endpoints: Damage to Human Health; Damage to Ecosystems; and Damage to Resource Availability. For endpoints a manual for Table 4.6 Midpoint impact categories included in LIME. Global Warming Ozone Layer Depletion Human Toxicity Ecotoxicity Photochemical Oxidant Acidification Eutrophication Urban Air Pollution Land Use Resource Consumption Waste Table 4.7 Midpoint impact categories modeled in ReCiPe. Climate Change Ozone Depletion Terrestrial Acidification Freshwater Eutrophication Marine Eutrophication Human Toxicity Photochemical Oxidant Formation Particulate Matter Formation Terrestrial Ecotoxicity Freshwater Ecotoxicity Marine Ecotoxicity Ionising Radiation Agricultural Land Occupation Urban Land Occupation Natural Land Transformation Water Depletion Mineral Resource Depletion Fossil Fuel Depletion 86 LIFE CYCLE ASSESSMENT HANDBOOK Table 4.8 Midpoint impact categories modeled in TRACI. Global warming Ozone Depletion Acidification Eutrophication Smog Formation Ecotoxicity Human Health: -criteria-related -cancer -noncancer Fossil Fuel Use Habitat/T&E Species Water Use T&E: Threatened & Endangered. panel weighting is available, but no operational generic weighting set have been developed. For the midpoints a monetisation method on the basis of pre- vention costs is provided. Reference for ReCiPe: Goedkoop M.J., Heijungs R, Huijbregts Mv De Schryver A., Struijs J., and 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; 6 January 2009. The Tool for the Reduction and Assessment of Chemical and other environ- mental Impacts (TRACI) http://cfpub.epa.gov/si/si_public_record_report.cfm?dirEntryId=103924 TRACI was developed by the US EPA. The impact categories are character- ised at the midpoint level, to concur with a higher level of societal agreement concerning the certainties of modeling within the cause-effect chain. TRACI was originally designed to assess chemical risks but has found usefulness in LCA. The methodologies underlying TRACI reflect state-of-the-art develop- ment and best-available practice for US conditions. A normalization database consistent with TRACI's impact categories and inventory flows is available, but no weighting is recommended in the models. Reference for TRACI: Bare, J.C., G.A. Norris, D.W. Pennington and T. McKone (2003) TRACI: The Tool for the Reduction and Assessment of Chemical and Other Environmental Impacts J Ind Ecol. 6(3): 49-78. 4.5 Variability in the LCIA Models The diversity offered in the various LCIA models that are available has created confusion, followed by criticism, of the use of LCIA, and LCA in general, since having several methodologies to choose from has the potential to generate dif- ferent results. The ISO 14042 standard on impact assessment published in 1999, LIFE CYCLE IMPACT ASSESSMENT 87 now part of ISO 14044 (2006), brought some standardization on basic principles. However, the ISO process did not provide detailed standardization (allowing for a "flexible standard/'). This flexibility allows for many different LCIA meth- ods to be ISO compliant (Pennington, Potting et ah 2004). Table 4.9 reports the main LCIA methodologies organized by their original school of appurtenance. Early SETAC working groups, later followed by UNEP/SETAC task forces, developed recommended best practice resulting in a relatively broad consen- sus on the best approaches, the underlying principles, and in some cases the models (see, for example, Udo de Haes et ah 2002). But these efforts fell short of getting the community to agree on a uniform, globallccepted set of LCIA meth- ods. Out of these deliberations and meetings, some agreement was reached. The most promising results are the following: • Consensus on the need to merge midpoint and endpoint models in a consistent framework to combine the advantages of both concepts (Bare et ah 2000). For example, midpoint indicators for climate change, in terms of C02-equivalents, and endpoint indica- tors, in terms of impacts on ecosystems and human health, in one consistent framework. • Development of guidance and a generic set of quality criteria for assessing good characterization modeling practices in LCIA and further its development (Udo de Haes et ah 2002, Margni et ah 2008). Table 4.9 Orientation of Main LCIA Methodologies. Distance-to-Target Critical Volumina Ecoscarcity (15) To Midpoint CMU9+) EDIP (9) TRACI (12) ILCD Handbook«» (15) To Damage or AoP EPS (5) Eco-indicator 99 (3) ILCD Handbook^ (3) Midpoint-Damage IMPACT 2002+ (14-4) LIME (11-4) ReCiPe(b) (18-3) IMPACT World+(c) (30-3) Numbers in parentheses (n) indicate the number of indicator categories. (a) Midpoint and Damage impact categories are proposed, but not integrated in a consistent framework. w Created from a methodological update of CML 2001 and Eco-indicator 99. (c) Created from a methodological update of IMPACT 2002+, LUCAS and EDIP. 88 LIFE CYCLE ASSESSMENT HANDBOOK • A global consensus among developers for fate, exposure and effect characterization of human toxicity and ecotoxicity with USEtox™ (Hauschild et ah 2008, Rosenbaum et al 2008). Building on these outcomes, more recently the Joint Research Center of the European Commission (EC-JRC) published in 2011 a guidance document pro- viding recommendations of best practice characterization framework, models and factors (for the common impact categories) that should be used for impact assessment in applications such as LCA (EC-JRC 2011). The ILCD Handbook is meant to support consistent and reliable business and policy instruments within the European Union (EU) related to products, natural resources, and waste management and their implementation, such as eco-labelling, eco- design, and green procurement. More recently, several initiatives have been started by government or by industry with the aim to harmonize LCA practices, particularly for environ- mental product declarations. To mention a few of them: Grenelle de l'environnement is a group (it could be called a "roundtable" in English) that brings together government, local authorities, trade unions, busi- ness and voluntary sectors to address and act on environmental issues. The Sustainability Consortium, launched in 2009, is an organization of diverse global participants working to make the world more sustain- able through better products, services and consumption. The efforts within the Consortium include the development of integrated tools that aim to improve informed decision making for product sustainability throughout the entire product lifecycle across all relevant consumer goods sectors. The Consumer Goods Forum works to harmonize framework and measure- ment systems in order to support better and more informed decision making. Among several recommendations, these stakeholder groups agreed on metrics and indicators to be applied with the aim to reduce the level of freedom when performing and communicating LCA results. Most of the recommendations made by the above mentioned initiatives build upon the recommendations made by the ILCD Handbook. In September 2011, the Forum announced release of the Global Protocol on Packaging Sustainability (GPPS) 2.0 to enable the consumer goods industry to better assess the relative sustainability of packaging, http://globalpackaging.mycgforum.com/ Several other methods for characterization models and factors are found in the literature, but without necessarily being integrated into a comprehensive LCIA methodology. A more detailed and extensive overview can be found in the ILCD Handbook (EC-JRC 2010b). 4.6 State-of-the-Art LCIA The ISO 14044 standard recommends that "the impact categories, category indicatorsand characterization models should be internationally accepted, i.e. based on an international agreement or approved by a competent international LIFE CYCLE IMPACT ASSESSMENT 89 body." Although this is not a mandatory "shall" requirement, only a few char- acterization models and factors currently satisfy this recommendation: the IPCC (Intergovernmental Panel on Climate Change) model calculating the Global Warming Potential (GWP) for the climate change midpoint impact category (Forster et al 2007), the WGMO (World and Global Meteorological Organization) model calculating the Ozone Depletion Potential (ODP) for the stratospheric ozone depletion impact category. For other impact categories, there has been only modest activity in international harmonization and scientific consen- sus (Hauschild, Goedkoop et ah 2012). This, however, is recently changing as researchers strive to find consensus in the models and underlying data. A breakthrough in identifying and recommending state-of-the art character- ization models and factors has been made by the European Platform on LCA. An extensive evaluation of existing LCIA methods and characterization mod- els was performed by model developers and scientific experts with the aim to identify the best existing practice. Through a consultation process involving listening to domain experts as well as stakeholders, the evaluation formed the basis of recommendations of characterization models and factors for impact categories at midpoint and at endpoint level (EC-JRC 2010b). The evaluation process was conducted in three steps. In the first step, the different characterization models used by each LCIA methodology in the characterization of impact categories and areas of protection were identified. This resulted in the identification of 156 characterization models stemming from eleven LCIA methodologies. In addition there were a few models which are not part of formal LCIA methodologies but showed interesting features. Of these, 91 were pre-selected and included in the following analysis. In the second step, criteria and procedures for the evaluation of character- ization models addressing midpoint and endpoint levels were developed. Five scientific criteria (completeness of scope, environmental relevance, scientific robustness and certainty, documentation & transparency & reproducibility, and applicability) and a stakeholder acceptance criterion were developed to evaluate all impact categories at the midpoint level and at the endpoint level. Each of these criteria was further detailed into a set of sub-criteria. Many sub-criteria were general and applied to each impact category. But for the scientific criteria on environmental relevance and scientific robustness and certainty, the sub criteria were developed specifically for each impact category, reflecting the central characteristics of the underlying impact pathway. An analysis of the impact pathway of each category helped identify key processes or aspects that should be considered in the characterization modeling, and these were the basis of formulating the category-specific sub criteria (flow sheets for each impact category can be found in EC-JRC 2010b). In the third step, the 91 shortlisted characterization models were further analyzed and compared to each impact category. The quality of the selected characterization model was assessed along three levels of recommendation: I - Recommended and satisfactory; II - Recommended but in need of some improvement; and III - Recommended, but to be applied with caution (See Table 4.10). T ab le 4 .1 0a B es t a va ila bl e ch ar ac te ri za tio n m od el s to m id po in t. M od el s th at a re c la ss if ie d as le ve l I , I I o r II I a re r ec om m en de d un de r th e IL C D (E C -J R C 2 01 1b ). Im pa ct C at eg or y C li m at e C h an ge O zo n e D ep le ti on H u m an T ox ic it y, C an ce r E ff ec ts H u m an T ox ic it y, N on -c an ce r E ff ec ts P ar ti cu la te M at te r/ R es pi ra to ry In or ga n ic s Io n is in g R ad ia ti on , H u m an H ea lt h Io n is in g R ad ia ti on , E co sy st em s P h ot oc h em ic al O zo n e F or m at io n B es t A m on g E xi st in g C ha ra ct er iz at io n M od el s B as el in e m od el o f 10 0 ye ar s of t he I P C C (F or st er e t a l., 2 00 7) S te ad y- st at e O D P s fr om t he W M O as se ss m en t (M on tz ka a n d F ra ze r, 1 99 9) U S E to x™ m od el ( R os en ba um e t a l. 20 08 ) U SE to x™ m od el ( R os en ba um e i a l. 20 08 ) R is kP ol l m od el ( R ab l an d S pa da ro 2 00 4) an d (G re co e t a l. 20 07 ) H u m an h ea lt h ef fe ct m od el a s de ve lo pe d by D re ic er e t a l. 19 95 ( re f. F ri sc hk ne ch t et a l 20 00 ) S cr ee ni ng L ev el E co lo gi ca l R is k A ss es sm en t (G ar ni er -L ap la ce e t a l., 2 00 8) ba se d on A M I m od el f ro m P ay et , 20 04 L O T O S -E U R O S as a pp li ed i n R eC iP e (V an Z el m e t a l. 20 08 ) In d ic at or R ad ia ti ve f or ci ng a s G lo ba l W ar m in g P ot en ti al ( G W P 10 0) O zo ne D ep le ti on P ot en ti al ( O D P ) C om pa ra ti ve T ox ic U ni t fo r h u m an s (C T U h) C om pa ra ti ve T ox ic U ni t fo r h u m an s (C T U h) In ta ke f ra ct io n fo r fi ne p ar ti cl es ( kg P M 2. 5- eq /k g) H u m an e xp os ur e ef fi ci en cy r el at iv e to U 23 5 C om pa ra ti ve T ox ic U ni t fo r ec os ys te m s (C T U e) T ro po sp he ri c oz on e co nc en tr at io n in cr ea se C la ss if ic at io n I I ΙΙ /Ι Π Π /Ι ΙΙ Ι/ Π II In te ri m II O W n n r< w > C D C /5 W on en w Z H X > o CO O o A ci di fi ca ti on E ut ro ph ic at io n, te rr es tr ia l E ut ro ph ic at io n, A qu at ic E co to xi ci ty , F re sh w at er L an d us e R es ou rc e D ep le ti on , W at er R es ou rc e de pl et io n, M in er al a nd F os si l A cc um ul at ed E xc ee da nc e (S ep pä lä e t a l 20 06 , P os ch e t a l 20 08 ) A cc um ul at ed E xc ee da nc e (S ep pä lä e t a l. 20 06 , P os ch e t a l 20 08 ) E U T R E N D m od el a s im pl em en te d in R eC iP e (S tru ijs e t a l, 20 09 b) U SE to x™ m od el , ( R os en ba um e t a l 20 08 ) M od el b as ed o n So il O rg an ic M at te r (S O M ) (M ilä i C an al s et a l 20 07 a) M od el fo r w at er c on su m pt io n as in t he Sw is s E co sc ar ci ty ( Fr is ch kn ec ht e t al , 20 08 ) C M L 2 00 2 (G ui ne e et a l 20 02 ) A cc um ul at ed E xc ee da nc e (A E) A cc um ul at ed E xc ee da nc e (A E) R es id en ce ti m e of n ut ri en ts i n fr es hw at er (P ) o r m ar in e en d co m pa rt m en t (N ) C om pa ra ti ve T ox ic U ni t f or e co sy st em s (C T U e) So il O rg an ic M at te r W at er u se r el at ed t o lo ca l s ca rc ity o f w at er Sc ar ci ty II 1 II II II /I II II I II II I - R ec om m en de d an d sa tis fa ct or y; I I - R ec om m en de d bu t in n ee d of s om e im pr ov em en t; II I - R ec om m en de d, b ut t o be a pp li ed w it h ca ut io n. A mix ed c la ss ifi ca tio n is re la te d to th e ap pl ic at io n of th e cl as si fie d m et ho d to d if fe re nt t yp es o f su bs ta nc es . \£> T ab le 4 .1 0b B es t a va ila bl e ch ar ac te ri za tio n m od el s fr om m id po in t to e nd po in t (H au sc hi ld , G oe dk oo p et a l. 20 12 ) Im pa ct C at eg or y C li m at e ch an ge O zo n e d ep le ti on H u m an T ox ic it y, C an ce r E ff ec ts H u m an T ox ic it y, N on -C an ce rE ff ec ts P ar ti cu la te M at te r/ R es pi ra to ry in or ga n ic s Io n is in g ra d ia ti on , h u m an h ea lt h Io n is in g ra d ia ti on , ec os ys te m s B es t A m on g E xi st in g C ha ra ct er iz at io n M od el s M od el d ev el op ed f or R eC iP e (D e S ch ry ve r an d G oe dk oo p, 2 00 9a ) M od el f or h u m an h ea lt h d am ag e de ve lo pe d fo r R eC iP e (S tr ui js e t a l, 20 10 ) D A L Y c al cu la ti on a pp li ed t o U S E to x™ m id po in t (A da pt ed f ro m H ui jb re gt s et a l, 20 05 ) D A L Y c al cu la ti on a pp li ed t o U S E to x™ m id po in t (A da pt ed f ro m H ui jb re gt s et a l, 20 05 ) A da pt ed D A L Y c al cu la ti on a pp li ed t o m id po in t (A da pt ed f ro m v an Z el m e t al , 20 08 , P op e et a l, 20 02 ) F ri sc hk ne ch t et a l, 20 00 In d ic at or D is ab il it y A dj us te d L if e Y ea rs ( D A L Y ) fo r H u m an H ea lt h P ot en ti al ly D is ap pe ar ed F ra ct io n of S pe ci es ( P D F m ^r ) fo r E co sy st em H ea lt h D is ab il it y A dj us te d L if e Y ea rs ( D A L Y ) D is ab il it y A dj us te d L if e Y ea rs ( D A L Y ) D is ab il it y A dj us te d L if e Y ea rs ( D A L Y ) D is ab il it y A dj us te d L if e Y ea rs ( D A L Y ) D is ab il it y A dj us te d L if e Y ea rs ( D A L Y ) C la ss if ic at io n in te ri m in te ri m II /i n te ri m in te ri m Ι/ Π in te ri m P h ot oc h em ic al o zo n e fo rm at io n A ci d if ic at io n E u tr op h ic at io n , T er re st ri al E u tr op h ic at io n , A q u at ic E co to xi ci ty L an d U se R es ou rc e D ep le ti on , W at er R es ou rc e D ep le ti on , M in er al a nd F os si l M od el f or d am ag e to h u m an h ea lt h as de ve lo pe d fo r R eC iP e (V an Z el m e t c d. , 20 08 ) M et ho d de ve lo pe d by v an Z el m e t a l. (2 00 7) a s in R eC iP e N o m et ho ds i de nt if ie d M od el f or d am ag e to e co sy st em ( fr es hw a- te r on ly ) S tr ui js e t a l, 20 09 b M od el f or s pe ci es d iv er si ty l os s as i n R eC iP e (D e S ch ry ve r an d G oe dk oo p, 20 09 b) M et ho d de ve lo pe d fo r R eC iP e (D e S ch ry ve r an d G oe dk oo p, 2 00 9b , G oe dk oo p an d D e S ch ry ve r, 2 00 9) D is ab il it y A dj us te d L if e Y ea rs ( D A L Y ) P ot en ti al ly d is ap pe ar ed f ra ct io n of p la nt sp ec ie s P ot en ti al ly D is ap pe ar ed F ra ct io n of S pe ci es ( P D F m ^r ) P ot en ti al ly D is ap pe ar ed F ra ct io n of S pe ci es ( P D F m 3 yr ) S ur pl us c os ts Π in te ri m in te ri m in te ri m in te ri m O nl y th re e m od el s ar e cl as si fie d ab ov e in te ri m , a nd o nl y th es e ar e re co m m en de d by th e IL C D H an db oo k. W Π n r< tu Π H > C D W QT > w H CO 94 LIFE CYCLE ASSESSMENT HANDBOOK If it was found that a 'best' existing characterization model could be identified, but this model was still not judged as being mature enough for recommendation at this time, it was then classified as "Interim." In cases where a 'best' could not be identified, either no model was recommended or it was classified as "interim." This did not mean the impact category was deemed irrelevant, but simply that more methodological development was needed before a recommendation or classification as interim could be made. The ILCD Handbook provides, therefore, an extensive analysis of the exist- ing characterization methods and recommendations for LCIA in the European context using reference year 2008. Since then methodological developments have continued, resulting in further advances of LCIA and others yet to come. 4.7 Future Development As shown in the previous section, LCA methodology has significantly devel- oped and matured over the last two decades. While gaining increasing acceptance, LCA still faces some major criticisms due to its holistic and inter- disciplinary character. Among these, current impact assessment methodolo- gies are not capable of (or are only partially capable of) consistently addressing the consequences of regional emissions (Udo de Haes, Finnveden et al. 2002; Potting and Hauschild 2006; von Klaus, Braune et at. 2007). Furthermore, they are still in their infancy in the development of some important resource-related impact categories such as water use, land use and in addressing issues such as spatial and temporal differentiation. 4.7.1 Spatially-Differentiated Assessment in LCIA In addition to global impact categories, such as global warming and ozone depletion, LCIA method developers recognize the need to have spatially- differentiated models for regional impact categories, due to the fact that differences in fate and exposure mechanisms and differences in sensitiv- ity and background levels for effect vary significantly depending on dif- ferent geographical contexts (Udo de Haes, Jolliet et al. 1999; Udo de Haes, Finnveden et al 2002). All LCIA approaches, IMPACT 2002+, ReCiPe, TRACI, LUCAS, LIME, etc., assume that the life cycle emissions are released in the geographical area where the methodology was been developed, i.e. in Europe, the US, Canada and Japan. This is an obvious and important limitation in LCIA methodology. Several research efforts have been attempting to develop spatially-differ- entiated characterization models and factors for current regional impact cat- egories (Potting and Hauschild 2006; Finnveden, Hauschild et al. 2009). Some LCIA methodologies such as EDIP (Hauschild and Potting 2005) and TRACI (Bare, Norris et al. 2003) also include a comprehensive set of regional impact categories allowing the practitioner to increase the discriminating power of LIFE CYCLE IMPACT ASSESSMENT 95 their LCA by assessing spatially-specific inventory emissions, if known. The geographical scope of the majority of these developments, however, remains restricted within a continental area and ignores transboundary emissions. Some of them provide a spatially-resolved assessment at the global scale, but usually by addressing a single impact category that is not available in a ready- to-use format for the LCA practitioner. More recently IMPACT World+ is being developed out of the need to offer a regionalized methodology at a global scale. Spatially-differentiated characterization models and factors have been developed for respiratory effects, toxic impacts, ionizing radiations, water use, acidification, eutrophication and land use impact categories, each of them based on an appropriate spatial scale. This latter was defined around the most sensitive modeling parameters, such as watersheds for water use impacts,biomes for land use impacts, or based on an archetype approach built upon the sensitive parameters. Particular attention has been given to the harmonization of modeling assumptions between different impact pathways. The uncertainty associated with the CFs for each of these "fine-scale" models has been deter- mined. These fine-scale CFs have been aggregated at the country, subconti- nental, and global scales using the geographical distribution of emissions (or emission proxis) as weighting factors. This resulted in CFs at different geo- graphical resolutions, each with its own associated uncertainty and spatial variability. The LC-IMPACT project, supported by the European Commission's 7th Framework Programme for Research and involving more than a dozen orga- nizations including a research center and an industry, represents an impor- tant initiative that demonstrates the growing interest and research activity around spatially-differentiated LCIA. One of the main objectives of this proj- ect includes the development of spatially-explicit CFs based on a global scale for land use, water exploitation, toxicants, priority air pollutants and nutrient (http://www.lc-impact.eu/about-lc-impact). The development toward a spatially-differentiated impact assessment is likely to modify the ISO paradigm. It is foreseeable in the near future that LCAs will be performed in three separate and successive phases: 1. Goal & Scope Definition, 2. Life Cycle Inventory, and 3. Impact Assessment, plus the interpretation step at each phase. In this way, a chemical emission or consump- tion of a given resource will have a different CF depending on the geographi- cal location where the environmental intervention occurred. So, instead of computing an LCI as a sum of chemical emissions over the whole life cycle; one will first characterize the emissions at each geographical location and then sum the impact scores among these locations. 4.7.2 Addressing Uncertainty and Variability in Characterization Factors Associating uncertainty information with CFs, encompassing both spatial variability and model uncertainty, is not current practice in LCA. So far, 96 LIFE CYCLE ASSESSMENT HANDBOOK existing LCIA methodologies only offer deterministic CFs without any addi- tional information as to their uncertainty. Recently, a few researchers have published papers that to provide methods to assess uncertainty propagation LCI and LCIA. A project within the UNEP/SETAC Life Cycle Initiative aims to establish recommended practice in uncertainty assessment and estimation within LCA and elaborate on guidance for practitioners and method develop- ers on estimation, communication, interpretation, and management of uncer- tainty in both LCI and LCIA. (http://lcinitiative.unep.fr/sites/lcinit/default. asp?site=lcinit&page_id=B70F576C-23B9-4D5F-9D87-6CA59AE3E0E6). The next generation of LCIA methodologies will systematically include uncertainty information associated with their CFs. For example, based on the analytical uncertainty propagation method developed by (Hong, Shaked et ah 2010), IMPACT World + already proposes uncertainty information associated with the CFs of impact categories. In addition, to address various sources of uncertainty, it also determines the uncertainty related to an unknown location of an emission by associating the corresponding geographical variability to each CF at a given geographical scale. Parallel initiatives, such as the cited LC-IMPACT project are also putting significant amount of research effort into providing quantitative information on various sources of uncertainty in life cycle impact assessment methods and corresponding factors. 4.7.3 Improving the Characterization of Resources Since the development of LCA in the early 1990s, impacts from resource use have been an integral part of LCA (Udo de Haes 2006). However, their evaluation and quantification of potential impacts remain one of the most debated issues in LCA methodology. Abiotic natural resources use (mineral and energy carrier) is already assessed by a wide variety of methods, how- ever, none were considered mature enough to be recommended for use in LCA (EC-JRC 2011b). A distinction is generally made between biotic and abi- otic natural resources. Although both are generally considered to be equally important, biotic resources have not received as much attention (Finnveden et ah 2009). Among the abiotic resources, the assessment of potential impacts related to water use and land use are still in their infancy, although two initiatives under the aegis of the UNEP/SETAC Life Cycle Initiative are raising interest and research activities around these issues (see WULCA (Water Use in Life Cycle Assessment) and LULCIA (Land Use in Life Cycle Impact Assessment) proj- ects at http://lcinitiative.unep.fr/). Depletion of minerals and fossil fuels are, nevertheless, addressed by several approaches that can be grouped into three categories (Finnveden, Hauschild et ah 2009; EC-JRC 2010): (i) Methods based on and an inherent property of the material such as exergy consumption or entropy production (Finnveden and Ostlund 1997; Bosch, Hellweg et ah 2007); LIFE CYCLE IMPACT ASSESSMENT 97 (ii) Methods addressing the scarcity of the resource by basing the assessment on the ratio between what is currently extracted related to some measure of available (EDIP) resources or reserves (CML); and (iii) Methods based on environmental impacts from future extrac- tions results in the need for additional efforts which can be trans- lated into higher energy or costs, and thus leads to an increased impact on the environment and economy (Müller-Wenk, 1998; Steen, 2006). Methods of this latter category are typically implemented in Eco-indicator 99, EPS, LIME, and IMPACT 2002+. The scopes of these approaches are so diverse that choosing one or the other might lead to completely different results and conclusions. Moreover, LCA practitioners and decisions makers are often not aware of what exactly these indicators represent and their underlying assump- tions/limitations. For example category (i) methods, although being relatively robust, are of little environmental relevance in expressing resource deple- tion (EC-JRC 2010). Category (ii) methods express the rate of disappearance of a given resource. When summing up these rates among different resource extractions over the life cycle of a product to calculate the impact score of this impact category, one implies the assumption that each resource is interchange- able (i.e. one can replace another). Although this may be true in some cases, it is doubtful this is always the case. For example, applying this implicit weight- ing makes the assumption that the depletion of 1kg of mineral x can be solved by using 1kg of mineral y with a lower disappearance rate independent of its functionality. Finally, several authors suggest that it is debatable to consider category (iii) methods based on environmental impacts from future extractions being part of the impact assessment, but should be included in the inventory analysis (Weidema, Finnveden et ah 2005; Finnveden, Hauschild et al 2009). No method has yet been able to follow the recommendations to move to a functionality-driven assessment framework as suggested by some research- ers (Jolliet, Müller-Wenk et al 2004; Margni, Gloria et al 2008). In such a frame- work, resources are considered to have only a functional value to humans and ecosystems, but no intrinsic value (i.e. a value for the sake of its existence as is the case for humans and ecosystems). This means that resource consumption has an impact only when its functionalities to humans and the ecosystems are degraded or lost. 4.7.4 Integrating Water Use and Consumption in LCIA The emergence of such of a framework based on resource functionality dis- sipation and degradation that accounts for competition between the users of a givenresource and eventually their adaptation capacity has, however, been observed by several researchers performing work in the context of the assessing water use, and more particularly by the framework developed by 98 LIFE CYCLE ASSESSMENT HANDBOOK consensus within WULCA (the UNEP/SETAC life cycle initiative working group focusing on water use impacts (Bayart, Bulle et ah 2010). In this frame- work water consumption leads to a modification of resource availability. The subsequent increased competition is captured by a competition index corre- lated with water scarcity (withdrawal or consumption to availability ratio). The first set of impact pathways addresses competition over freshwater resources between different human activities due to an insufficient supply of the resource. A reduction in water availability to humans can potentially affect human health if water were fulfilling human essential needs (domestic use, agriculture or aquaculture). If financial resources are available, there can be adaptations by using a functionally-equivalent alternative that may, in turn, shift the environmental burdens to other life cycle stages and impact catego- ries. Direct impacts on human health related to water deprivation have been addressed by (Pfister, Koehler et ah 2009; Motoshita, Itsubo et ah 2010; Boulay, Bulle et ah 2011a). The indirect impact pathways consider that nobody suffers from water deprivation in world wealth regions as competing users have the capacity to adapt to a reduced water availability (for example, by desalinating water or by importing food). Indirect impacts generated by such compensation scenarios are under development (Boulay, Bulle et ah 2011b). The second set of impact pathways relates to insufficient freshwater for existing ecosystems due to increased withdrawal by humans. Several method- ologies have been published recently to address different impact pathways by linking water use with impacts on ecosystem quality: decreased terrestrial bio- diversity due to water consumption (Pfister, Koehler et ah 2009); disappearance of terrestrial plant species due to shallow ground water withdrawal and related lowering of the water table (Zelm, Schipper et ah 2010); and the effects of water consumption on freshwater fish species (Hanafiah, Xenopoulos et ah 2011). The third pathway has, so far, received less attention. This pathway origi- nally addressed the reduced availability of freshwater for future generations and outlines long-term depletion. Only Pfister, Koehler et ah (2009) have attempted to quantify the impact on future freshwater availability through a backup-technology approach to evaluate the impact of water consump- tion above their renewability rate. However, it is expected that this area of protection will be addressed in the near future in the interest of ecosystem services (i.e. addressing the functional value as mentioned above), where reduced ecosystems services will have a direct impact on human society which is measurable through economic consequences. 4.7.5 Resources and Ecosystem Services Areas of Protection People everywhere rely on ecosystems and the services they provide. The loss or degradation of ecosystem services will have severe impacts on human well-being and have a profound effect on businesses. Higher operating costs or reduced operating flexibility should be expected due to diminished or degraded resources (such as freshwater) or increased regulation (MA 2005). Moving towards the definition of an Area of Protection that evaluates the LIFE CYCLE IMPACT ASSESSMENT 99 impact of human activities on global ecosystem integrity and ecosystem ser- vices sustainability certainly answer an increasing interest among a variety of stakeholders along the production and consumption value chain, looking at a comprehensive view of the direct and indirect impacts generated by their product and services both for humans and ecosystems. 4.7.6 Expanding Land Use Burdens on Biodiversity in Ecosystem Services Regarding biodiversity, impacts solely related to terrestrial biodiversity (PDF. m2.year) have been implemented in LCIA methodologies. They are too restric- tive on their spatial coverage in that they are generally limited to the European continent and fail to address particular ecosystems when it comes to other countries. Or, they are too restrictive in the impact pathways they cover. In order to fill in methodological gaps and to answer the need for integrating and harmonizing impact indicators as extensively justified in the literature (Müller- Wenk 1998; Lindeijer, Müller-Wenk et al 2002; Milä i Canals, Bauer et al 2007) a working group within the UNEP/SETAC Life Cycle Initiative (LULCIA) pro- pose a guideline to build methods for land use impact assessment (Koellner et al 2012) refining the principles that have already been proposed by others (Mila i Canals, Bauer et al 2007) who recommend model developers address the calculation of land use interventions and land use impacts, the issue of impact reversibility, the spatial and temporal distribution of such impacts, and the assessment of absolute or relative ecosystem quality changes. This method relates land use to six new indicators in addition to biodiversity: biotic production (BPP), erosion regulation (ERP), freshwater regulation (FWRP), mechanical and physicochemical water purification (MWPP and PCWPP), and carbon sequestration (CSP) potentials, which represent provision and regulation ecosystem services, as defined in the Millennium Assessment (MA 2005). These indicators of land use impacts are calculated as the product of surface occupied (or transformed), occupation (or transformation) time, and a parameter describ- ing the land quality (or ecosystem functionality) loss. It is noteworthy that the ecosystem services approach adopted for land use impact assessment is quite similar to the functional equivalency approach adopted for water use. References Bare, J., Hofstetter, P., et al (2000). "Life Cycle Impact Assessment Workshop Summary - Midpoints versus Endpoints: The Sacrifices and Benefits." International Journal of Life Cycle Assessment 5(6): 318-326. Bare, J.C., Norris, G.A., et al (2003). "TRACI: The Tool for the Reduction and Assessment of Chemical and Other Environmental Impacts." Journal of Industrial Ecology 6(3-4): 49-78. Bare, J.C. and Gloria, T.R (2006). "Critical analysis of the mathematical relationships and compre- hensiveness of life cycle impact assessment approaches." Environmental Science & Technology 40(4): 1104-1113. 100 LIFE CYCLE ASSESSMENT HANDBOOK Bayart, J.B., Bulle, C, et al. (2010). "A framework for assessing off-stream freshwater use in LCA." International Journal of Life Cycle Assessment 15(5): 439-453. Bosch, M., Hellweg, S., et al. (2007). "Applying cumulative exergy demand (CExD) indicators to the ecoinvent database." The International Journal of Life Cycle Assessment 12(3): 181-190. Boulay, A.-M., Bulle, C, et al. (2011a). "Regional Characterization of Freshwater Use in LCA: Modeling Direct Impacts on Human Health." Environmental Science & Technology 45(20): 8948-8957. Boulay, A.-M., Bulle, C, et al. (2011b). "LCA Characterisation of Freshwater Use on Human Health and Through Compensation." Towards Life Cycle Sustainability Management. M. Finkbeiner, Springer Netherlands: 193-204. Brand, G., Braunschweig, A., et al. (1997). "Weighting in Ecobalances with the Ecoscarcity Method - Ecofactors;" Environment Series No. 297. Bern, Switzerland, Swiss Agency for the Environment, Forests, and Landscape (SAEFL). BUS (1984). Ökobilanzen von Packstoffen, Schriftenreihe Umweltschutz Nr.24. Bern, Switzerland, Bundesamt für Umweltschutz. De Schryver, A. and Goedkoop, M. (2009a). "Climate Change." Chapter 3 in: Goedkoop, M., Heijungs, R., Huijbregts, M.A.J., De Schryver, A., Struijs, J., and Van Zelm, R. (2009). ReCiPe 2008 A life cycle impact assessment method whichcomprises harmonised category indicators at the midpoint and the endpoint level. Report I: Characterisation factors, first edition. De Schryver, A. and Goedkoop, M. (2009b). "Mineral Resource." Chapter 12 in: Goedkoop, M., Heijungs, R., Huijbregts, M.A.J., De Schryver, A., Struijs, J., and Van Zelm, R. (2009). ReCiPe 2008. A life cycle impact assessment method which comprises harmonised category indicators at the midpoint and the endpoint level. Report I: Characterisation factors, first edition. Dreicer, M., Tort, V, and Manen, P. (1995). ExternE, Externalities of Energy, Vol. 5 Nuclear, Centr d'etude sur l'Evaluation de la Protection dans le domaine nucleaire (CEPN), edited by the European Commission DGXII, Science, Research and development JOULE, Luxembourg. EC-JRC (2010a). "Framework and requirements for Life Cycle Impact Assessment (LCIA) mod- els and indicators." ILCD Handbook - International Reference Life Cycle Data System, European Commission - Joint Research Center. EC-JRC (2010b). "An analysis of existing Environmental Impact Assessment methodologies for use in Life Cycle Assessment - Background document." ILCD Handbook - International Reference Life Cycle Data System, European Commission - Joint Research Center. EC-JRC (2010c). "General Guide for Life Cycle Assessment - Detailed Practice." ILCD Handbook - International Reference Life Cycle Data System, European Commission - Joint Research Center. EC-JRC (2011). "Recommendations for LCIA in the European context - based on existing envi- ronmental impact assessment models and factors." ILCD Handbook - International Reference Life Cycle Data System. ISPRA, European Commission - Joint Research Center. EPA (2004). An Examination of EPA Risk Assessment Principles and Practices, EPA/100/B-04/00. US Environmental Protection Agency, Office of the Science Advisor. Washington, DC. EPA (2006). Life Cycle Assessment: Principles and Practice, EPA/600/R-06/060. US Environmental Protection Agency, Office of Research & Development. Cincinnati, Ohio. Finnveden, G., Hauschild, M.Z., et al. (2009). "Recent developments in Life Cycle Assessment." Journal of Environmental Management 91(1): 1-21. Finnveden, G. and Ostlund, P. (1997). "Exergies of natural resources in life cycle assessment and other applications." Energy 22: 923-931 Forster, P., Ramaswamy, V, Artaxo, P., Berntsen, T., Betts, R., Fahey, D.W., Haywood, J., Lean, J., Lowe, D.C., Myhre, G., Nganga, J., Prinn, R., Raga, G., Schulz, M. and Van Dorland, R. (2007). "Changes in Atmospheric Constituents and in Radiative Forcing." In: Climate Change 2007: The Physical Science Basis IPCC 2O07.Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Frischknecht, R., Steiner, R., and Jungbluth, N. (2008). "Methode der ökologischen Knappheit - Ökofaktoren 2006," ö.b.u. und Bundesamt für Umwelt, Bern. LIFE CYCLE IMPACT ASSESSMENT 101 Frischknecht, R., Braunschweig, A., Hofstetter, P., and Suter, P. (2000). "Modelling human health effects of radioactive releases in Life Cycle Impact Assessment." Environmental Impact Assessment Review, 20 (2) pp. 159-189. Garnier-Laplace, }.C, Beaugelin-Seiller, K., Gilbin, R., Delia-Vedova, C, Jolliet, O., and Payet, J. (2008). "A Screening Level Ecological Risk Assessment and ranking method for liquid radioactive and chemical mixtures released by nuclear facilities under normal operating conditions." Proceedings of the International conference on radioecology and environmental protec- tion, 15-20 June 2008, Bergen. Goedkoop and De Schryver (2009). "Fossil Resource." Chapter 13 in: Goedkoop, M., Heijungs, R., Huijbregts, M.A.J., De Schryver, A., Struijs, J., and van Zelm, R. (2009). ReCiPe 2008 - A life cycle impact assessment method which comprises harmonised category indicators at the midpoint and the endpoint level. Report I: Characterisation factors, first edition. Greco, S.L., Wilson, A.M., Spengler, J.D., and Levy, J.I. (2007). "Spatial patterns of mobile source particulate matter emissions-to-exposure relationships across the United States." Atmospheric Environment (41), 1011-1025. Guinee, J. B., Gorree, M., et al. (2002). Handbook on Life Cycle Assessment. Operational Guide to the ISO standards. I: LCA in Perspective. Ila: Guide. lib: Operational Annex. Ill: Scientific background. Dordrecht,, Kluwer Academic Publishers. Hanafiah, M.M., Xenopoulos, M. A., et al. (2011). "Characterization Factors for Water Consumption and Greenhouse Gas Emissions Based on Freshwater Fish Species Extinction." Environmental Science & Technology 45(12): 5272-5278. Hauschild, M., Goedkoop, M., et al. (2012). "Best existing practice for characterization mod- elling in Life Cycle Impact Assessment." International Journal of Life Cycle Assessment submitted. Hauschild, M. and Potting, J. (2005). "Spatial differentiation in Life Cycle Impact Assessment: The EDIP03 Methodology." Environmental News No. 80. Guidelines from the Danish Environmental Protection Agency, Copenhagen, Denmark. Hauschild, M.Z., Huijbregts, M.A.J., Jolliet, O., MacLeod, M., Margni, M., van de Meent, D., Rosenbaum, R.K., and McKone, T.E. (2008). "Building a model based on scientific consen- sus for life cycle impact assessment of chemicals: the search for harmony and parsimony." Environmental Science & Technology 42(19): 7032-7037. Heijungs, R., Guinee, J., et al. (1992). Environmental Life Cycle Assessment of Products: Guide and Background. CML. Leiden, The Netherlands. Hong, J., Shaked, S., et al. (2010). "Analytical uncertainty propagation in life cycle inventory and impact assessment: application to an automobile front panel." The International Journal of Life Cycle Assessment 15(5): 499-510. Huijbregts, M.A.J., Rombouts, L.J.A., Ragas, A.M.J., and Van de Meent, D. (2005). "Human- toxicological effect and damage factors of carcinogenic and noncarcinogenic chemicals for life cycle impact assessment." Integrated Environ. Assess. Manag. 1:181-244. ISO (2006a). Environmental Management - Life Cycle Assessment - Principles and Framework. Brussels, International Standards Organization. ISO 14040. ISO (2006b). Environmental Management - Life Cycle Assessment - Requirements and Guidelines,. Brussels, International Standards Organization. ISO 14044. Jolliet, O., Müller-Wenk, R., et al. (2004). "The LCIA Midpoint-damage Framework of the UNEP/ SETAC Life Cycle Initiative." International Journal of Life Cycle Assessment 9(6): 394-404. Lindeijer, E., Müller-Wenk, R., et al. (2002). "Impact Assessment of Resources and Land Use." Life-Cycle Impact Assessment: Striving towards Best Practice. H. Udo de Haes, G. Finnveden, M. Goedkoopef al. Pensacola, USA, Society of Environmental Toxicology and Chemistry (SETAC): 11-64. MA (2005). Millennium Ecosystem Assessment (MA). Ecosystems and Human Well-Being: Synthesis. Island Press, Washington. 155pp. Margni, M., Gloria, T., et al. (2008). Guidance on how to move from current practice to recommended practice in Life Cycle Impact Assessment, UNEP-SETAC Life Cycle Initiative. 102 LIFE CYCLE ASSESSMENT HANDBOOK Milä i Canals, L., Bauer, C , et al. (2007). "Key Elements in a Framework for Land Use Impact Assessment within LCA." International Journal of Life Cycle Assessment 12(1): 5-15. Milä i Canals, L., Romanya, J., and Cowell, S.J. (2007). "Method for assessing impacts on life support functions (LSF) related to the use of 'fertile land' in Life Cycle Assessment (LCA)." Journal of Cleaner Production 15 1426-1440. Montzka, S.A., Butler, J.H., Elkins, J.W., Thompson, T.M., Clarke, A.D., and Lock, L.T. (1999). "Present and future trends in the atmospheric burden of ozone-depleting halogens." Nature 398: 690-694. Motoshita, M., Itsubo, N., et al. (2010). "Development of impact factors on damage to health by infectious diseases caused by domestic water scarcity." The International Journal of Life Cycle Assessment:1-9. Müller-Wenk, R. (1998). Land Use - The Main Threat to Species. How to Include Land Use in LCA. Switzerland, Universität of St.Gallen: 46. Payet, J. (2004). Assessing toxic impacts on aquatic ecosystems in LCA. Doctoral thesis 3112, Ecole Poly technique Federale de Lausanne (EPFL, CH-1015 Lausanne), pp. 214. Pennington, D., Potting, J., et al. (2004). "Life Cycle Assessment Part 2: Current Impact Assessment Practice." Environment International 30(5): 721-739. Pfister, S., Koehler, A., et al. (2009). "Assessing the Environmental Impacts of Freshwater Consumption in LCA." Environmental Science & Technology 43(11): 4098^4104. Pope, CA., Burnett, R.T., Thun, M.J., Calle, E.E., Krewski, D., Ito, K., and Thurston, G.D. (2002). "Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution." Journal of the American Medical Association 287,1132-1141. Posch, M., Seppälä, J., Hettelingh, J.P., Johansson, M., Margni M., and Jolliet, O. (2008). "The role of atmospheric dispersion models and ecosystem sensitivity in the determination of charac- terisation factors for acidifying and eutrophying emissions in LCIA." International Journal of Life Cycle Assessment (13) pp. 477-486. Potting, J. and Hauschild, M.Z. (2006). "Spatial differentiation in life cycle impact assessment - A decade of method development to increase the environmental realism of LCIA." International Journal of Life Cycle Assessment 11:11-13. Rabl, A. and Spadaro, J.V. (2004). The RiskPoll software, version is 1.051 (dated August 2004). www.arirabl.com. Rosenbaum, R.K., Bachmann, T.M., Gold, L.S., Huijbregts, M.A.J., Jolliet, O., Juraske, R., Köhler, A., Larsen, H.F., MacLeod, M., Margni, M., McKone, T.E., Payet, J., Schuhmacher, M., van de Meent, D., and Hauschild, M.Z. (2008). "USEtox - The UNEP/SETAC toxicity model: recom- mended characterisation factors for human toxicity and freshwater ecotoxicity in Life Cycle Impact Assessment." International Journal of Life Cycle Assessment 13(7): 532-546. Seppälä, J., Posch, M., Johansson, M., and Hettelingh, J.P (2006). "Country-dependent Characterisation Factors for Acidification and Terrestrial Eutrophication Based on Accumulated Exceedance as an Impact Category Indicator." International Journal of Life Cycle Assessment 11(6): 403^16. Steen, B. (1999a). "A systematic approach to environmental priority strategies in product devel- opment (EPS)." Version 2000 - General system characteristics, CPM Report. Chalmers University of Technology. Sweden. Steen, B.A. (2006). "Describing values in relation to choices in LCA." International Journal of Life Cycle Assessment 11(4): 277-283. Struijs, J., Beusen, A., van Jaarsveld, H., and Huijbregts, M.A.J. (2009). "Aquatic Eutrophication." Chapter 6 in: Goedkoop, M., Heijungs, R., Huijbregts, M.A.J., De Schryver, A., Struijs, J., and Van Zelm, R. (2009). ReCiPe 2008 A life cycle impact assessment method which comprises harmonised category indicators at the midpoint and the endpoint level. Report I: Characterisation factors, first edition. Struijs, J., van Dijk, A., Slaper, H., van Wijnen, HJ., Velders, G.J.M., Chaplin, G., and Huijbregts, M.A.J. (2010). "Spatial- and Time-Explicit Human Damage Modeling of Ozone Depleting Substances in Life Cycle Impact Assessment." Environmental Science & Technology 44 (1): 204-209. LIFE CYCLE IMPACT ASSESSMENT 103 Udo de Haes, H. (2006). "Sustainable Management of Natural Resources in an Life-cycle Perspective (Issue Editor: Helias A. Udo de Haes)." International Journal of Life Cycle Assessment 11(1): 2-2. Udo de Haes, H.A., Finnveden, G., et al (2002). Life-Cycle Impact Assessment: Striving towards Best Practice. Pensacola (US), Society of Environmental Toxicology and Chemistry (SETAC). Udo de Haes, H.A., Jolliet, O., et al. (1999). "Best Available Practice Regarding Impact Categories and Category Indicators in Life Cycle Impact Assessment - Background Document for the Second Working Group on Life Cycle Impact Assessment of SETAC-Europe (WIA-2)." International Journal of Life Cycle Assessment 4(2): 66-74. van Zelm, R., Huijbregts, M.A.J., Van Jaarsveld, H.A., Reinds, G.J., De Zwart, D., Struijs, J., and Van de Meent, D. (2007). "Time horizon dependent characterisation factors for acidifica- tion in life-cycle impact assessment based on the disappeared fraction of plant species in European forests." Environmental Science & Technology 41(3): 922-927. van Zelm, R., Huijbregts, M.A.J., Den Hollander, H.A., Van Jaarsveld, H.A., Sauter, F.J., Struijs, J., Van Wijnen, H.J., and Van de Meent, D. (2008). "European characterization factors for human health damage of PM10 and ozone in life cycle impact assessment." Atmospheric Environment 42,441-453. van Zelm, R., Schipper, A.M., et al. (2010). "Implementing Ground water Extraction in Life Cycle Impact Assessment: Characterization Factors Based on Plant Species Richness for the Netherlands." Environmental Science & Technology 45(2): 629-635. von Klaus, S., Braune, A., et al. (2007). "Spatial differentiation in LCA - Moving forward to more operational sustainability." Technikfolgenabschätzung - Theorie und Praxis 3(16). Weidema, B.P., Finnveden, G., et al. (2005). "Impacts from Resource Use A - common position paper." International Journal of Life Cycle Assessment 10(6): 382. 5 Sourcing Life Cycle Inventory Data Mary Ann Curran* US Environmental Protection Agency, Cincinnati, OH, USA Abstract The collection and validation of quality life cycle inventory (LCI) data can be the most difficult and time-consuming aspect of developing a life cycle assessment (LCA). Large amounts of process and production data are needed to complete the LCI. For many studies, the LCA analyst at some point will need to collect process data from original sources. In these instances, the increasing sophistication expected of LCA studies has tended to make the task of data collection more demanding, rather than less, as the scrutiny placed on published LCA results has grown. As a result, a growing number of public databases of LCI data are becoming available to provide data for the more common commodity processes and services. This chapter discusses historical and current practices in sourcing LCI data (other than using the data that come with commercial off the shelf (COTS) LCA tools). Data can be acquired by utilizing dedicated LCI databases, non-LCI databases, publica- tions, and sources, as well as by implementing estimation techniques that use eco- nomic input/output tables to generate LCI data where field measurements cannot be easily made. Finally, the chapter discusses emerging approaches for reporting process inventory data, including manufacturer self-reporting, using open-source models for data collection and annotation, and "crowdsourcing" of LCI data. Keywords: Life cycle assessment, life cycle inventory, data, database, input-output 5.1 Introduction As with any assessment tool, data is the driving force behind Life Cycle Assessment (LCA). Large amounts of process and production data are needed to complete the life cycle inventory (LCI). Raw material inputs, energy use, ratio of main product to co-products, production rates, and environmental releases must all be quantified for each process in the system. There are many ways to generate LCI data. Perhaps the easiest, but most costly way, is using a commercial off the shelf (COTS) software tool, such as * The views expressed in this chapter are those of the author and do not necessarily reflect the views or policies of the US Environmental Protection Agency. Mary Ann Curran (ed.) Life Cycle Assessment Handbook: A Guide for Environmentally Sustainable Products, (105-142) © 2012 Scrivener Publishing LLC 105 106 LIFE CYCLE ASSESSMENT HANDBOOK GaBi or SimaPro. Software, however, is the subject of another chapter and will not be covered here. Another option is to build the LCI ina tailor-made fashion, directly from data sources. This approach is covered in the following sections. In many instances, creating an LCI begins with the collection of raw data which are data extracted from various sources, such as bookkeeping of a plant, national statistics, technical journals, etc., but not yet related to the process for which the dataset is being developed. Typically, a number of sources are needed to be called upon to collect a sufficient amount of data. Other examples of data sources that may be drawn from or utilized include the following: • Meter readings from equipment • Equipment operating logs/journals • Industry data reports, databases, or consultants • Laboratory test results • Government documents, reports, databases, and clearinghouses • Other publicly available databases or clearinghouses • Journals, papers, books, and patents • Reference books • Trade associations • Related /previous life cycle inventory studies • Equipment and process specifications • Best engineering judgment (EPA 2006) Once raw data are collected, following a pre-determined data collec- tion approach, unit process datasets can be created by defining mathemati- cal relationships between the raw data and various flows associated with the dataset in a defined reference flow. Data modeling requirements, with desired quality attributes and adequate documentation, are specified to accu- rately transform raw data into unit process datasets, and incorporate proper review and documentation to address verification and transparency issues (Consoli, Allen et al. 1993; Curran 2011). Therefore, understanding how data flow from raw data providers to LCI data users (shown in Figure 5.1) is important because data move from the raw state to and through datasets and databases. Recycling provides an example of some of the strengths and limitations encountered in gathering data. For some products, economic-driven recycling has been practiced for many years, and infrastructure and markets for these materials already exist. Data are typically available for these products, includ- ing recycling rates, the consumers of the reclaimed materials, and the resource requirements and environmental releases from the recycling activities (col- lection and reprocessing). Data for materials currently at low recycling rates with newly forming recycling infrastructures are more difficult to obtain. In either case, often the best source for data on resource requirements and envi- ronmental releases is the processors themselves. For data on recycling rates and recycled material, consumers and processors may be helpful, but trade SOURCING LIFE CYCLE INVENTORY DATA 107 Data in the supply chain Raw data collection Raw data collection - extended Raw data collection - crowd sourcing Database management * Unit process database development System n nodellin 3 Networking of databases JL Storing of datasets (database) 31 Database review and validation t Aggregated process dataset development 4— User feedback * User interaction information: Algorithms for linking database and handling multi-output datasets Figure 5.1 Row of data from raw data through to LCI data user with feedback loops (UNEP/ SETAC 2011). associations as well as the consumers of the recycled materials can provide data. For materials that are recycled at low rates, data will be more difficult to find. Two other areas for data gathering relate to the system as a whole and to comparisons between and among systems. It is necessary to obtain data on the weights of each component in the product evaluated, either by obtaining product specifications from the manufacturer or by weighing each component. These data are then used to combine the individual components in the over- all system analysis. Equivalent use ratios for the products compared can be developed by surveying retailers and consumers, or by reviewing consumer or trade association periodicals. 5.2 Developing LCI to Meet the Goal of the Study 5.2,1 Considerations in Choosing Data Sources For each life cycle stage, unit process, or type of environmental release, it is important to specify the necessary data source and/or type required to pro- vide sufficient accuracy and quality to meet the study goal. Naturally, LCAs should use the most appropriate datasets and modeling approaches to meet the specific goal required to satisfactorily answer the question(s) posed for study. Therefore, in choosing between using readily available datasets or develop- ing data (from "scratch") it is important to have a clear understanding of the 108 LIFE CYCLE ASSESSMENT HANDBOOK study's goal and scope. In making such a choice, the following factors should be taken into consideration: • Physical delimitation of activities such as principal process type (e.g., from site-specific to industry-average type) and the specific size of the process to be modeled; • Impact categories to be evaluated during the impact assessment; • Technology covered; • Time period covered; • Geographical area covered; • Cut-off rules1 for data, if any, are applied (these rules should pro- vide a rationale for the significance of the various flows of the unit process dataset); • Provision of uncertainty information for inputs and outputs of the process to allow for uncertainty analysis; • Targeted databases for unit process datasets that are considered to be high priority; and • Intended use of the dataset in general (applications, modeling situations including attributional or consequential modeling, comparative assertions). A well-defined scope helps answer questions, which, in turn, help the ana- lyst determine the level or type of information that is required. For example, even when the analyst can obtain actual industry data, in what form and to what degree of specificity should the analyst show the data (i.e., the range of values observed, industry average, plant-specific data, best available con- trol techniques, etc.)? Recommended practice for external life-cycle inventory studies includes the provision of a measure of data variability in addition to averages. Frequently, the measure of variability will be a statistical parameter, such as standard deviation (EPA 2006). 5.2.2 A Word on Consequential Life Cycle Assessment LCA was initially developed to assess industrial systems related to consumer products. Since then, there has been a distinct shift in applying it to larger scales of industrial operations. By 2005, LCA practitioners began making a distinc- tion between how LCAs that accounted for stoichiometric-like relationships between physical flows to and from a product or process in an attributional style, to a ones that were more encompassing of the consequences of change in response to decisions, in a consequential LCA (Curran et ah, 2005). As a result, the process of system expansion (to avoid or deal with the allocation 1 Non-reference product flows, waste flows, and elementary flows that can safely be labeled as "irrelevant" can be ignored (i.e. "cut-off"). However, care must be taken to not cut off more flows and related impacts than are acceptable to still meet the goal and scope, and the datasets used to model a system meet the required completeness (UNEP/SETAC 2011). SOURCING LIFE CYCLE INVENTORY DATA 109 problem in multi-product systems) is an inherent part of consequential LCA studies. Ultimately, choosing between an attributional or a consequential LCA is decided by the defined goal of the study. The choice will also influence sys- tem boundaries related to how allocation is conducted as well as other meth- odological choices, such as the definition of functional unit and the choice of life cycle impact assessment (LCIA) methods (Finnveden 2009). The decision to use marginal data can be significant for modeling systems that include activities such as electricitygeneration, land use, etc. or other areas in which a change in the balance between supply and demand for a good or service can have a far-reaching impact. For example, Searchinger et al. found an attributional analysis of US corn-based ethanol resulted in a 20% decrease in greenhouse gas emissions compared to conventional gasoline. However, in a consequential analysis to account for policy-driven increases in output, they predicted a 47% increase in emissions compared to gasoline, due to land use changes induced by higher prices of corn, soybeans and other grains from anticipated additional demand for corn starch for ethanol production. A consequential LCA is conceptually complex because it includes addi- tional, economic concepts such as marginal production costs, elasticity of sup- ply and demand, etc. Consequential LCA depends on descriptions of economic relationships embedded in models. It generally attempts to reflect economic relationships by extrapolating historical trends in prices, consumption and outputs. Some of the models are also much less transparent than the linear and static model of attributional LCA. Their results can also be very sensitive to the built-in assumptions. All these add to the risk that inadequate assumptions or other errors significantly affect the final LCA results. To reduce this risk, it is important to ensure that the various results regarding different consequences can be explained using credible arguments. It is possible that the inventory results of a consequential LCA will be nega- tive, if the change in the level of production causes a reduction in emissions greater than the emissions from the production of the product. This does not mean that the absolute emissions from the production of the product are nega- tive, but that the production of the product will cause a reduction in emissions elsewhere in the system. In the end, both approaches are legitimate and fulfill different needs (Ekvall et al 2005). The distinction between attributional and consequential LCA is one example of how choices in the Goal and Scope Definition of an LCA should influence methodological and data choices for the LCI and LCIA phases. 5.3 Types of LCI Data Clearly, defining the required data sources and types prior to data collection helps to reduce costs and the time required to collect the data. Whenever possi- ble, it is best to get well-characterized industry data for production processes. Manufacturing processes can change over time by becoming more efficient, adopting newer technology, incorporating changes to emissions standards, 110 LIFE CYCLE ASSESSMENT HANDBOOK etc. Therefore, it is important to seek current data. Several categories of data are often used in inventories. Starting with the most disaggregated, these are: • Individual process- and facility-specific: data from a particular operation within a given facility that are not combined in any way. • Composite: data from the same operation or activity combined across locations. • Aggregated: data combining more than one process operation. • Industry-average: data derived from a representative sample of locations and believed to statistically describe the typical opera- tion across technologies. • Generic: data whose representativeness may be unknown but which are qualitatively descriptive of a process or technology. Data can be classified by how they are created: • Site-specific (directly measured or sampled) • Modeled, calculated or estimated • Non-site specific (i.e., surrogate data) • Non-LCI data (i.e., data not originally intended for use in an LCI) • Vendor data Data sources are either primary or secondary: 1. Primary data come directly from the source, including: • Interviews, • Questionnaires or surveys, • Bookkeeping or enterprise resource planning (ERP) system, • Data collection tools (online, offline), and • On-site measurements. 2. Secondary data come from reports found in: • Databases, • Statistics, and • Open literature. Unit process datasets are the basis of every LCI database and the foundation of all LCA applications. A unit process dataset is obtained as a result of quan- tifying inputs and outputs in relation to a quantitative reference flow from a specific process. These inputs and outputs are generated from mathematical relationships based on raw data that have not previously been related to the same reference flow. An aggregated process dataset is obtained from a collec- tion of similar unit process or other aggregated datasets. Most often, datasets are aggregated to protect business-sensitive, competition-sensitive, or propri- etary information, including trade secrets, patented processes, process infor- mation used to easily derive costs, etc. SOURCING LIFE CYCLE INVENTORY DATA 111 "Unit process" is defined as "smallest element considered in the life cycle inventory analysis" in ISO 14040 (ISO 2006). Unit process datasets are usually distinguished from aggregated process datasets. However, when used in creat- ing an LCI, an aggregated process dataset may be considered as representing a unit process. 1 Single operation Database Gate-to-gate Database Cradle-to-gate LCI Database 4 5 6 Partly Partial vertical Cradle-to-grave terminated aggregation LCI Database ^ B Database ^ B Database I Background 2 2" 2" 3 3' 3" Use End of life Background Background Background V V 2' 2" 3' 3" Use End of life Use End of life Use End of life Use End of life 7 Single operation Database Background to* 2 2' 2" 3 3' 3" Use End of life 8 Gate-to-gate Database Background Use End of life Cradle-to-gate LCI Database Background Background 1 r r Use End of life 10 11 12 Partly Partial vertical Cradle-to-grave terminated aggregation LCI Use End of life Use End of life Figure 5.2 Unit process datasets within databases can be aggregated multiple ways, including various combinations of horizontal and vertical aggregation (UNEP/SETAC 2011). 112 LIFE CYCLE ASSESSMENT HANDBOOK The required level of aggregated data should be specified (as guided by the study's goal), for example, whether data are representative of one process or of several processes. Figure 5.2 depicts the possible variations to aggre- gate processes (steps 2 through 11). Step 1 indicates no aggregation (a single process); step 12 is the complete cradle-to-grave LCI, the ultimate form of aggregation. 5.4 Private Industrial Data Complete and thorough inventories often require using proprietary data that are provided by either the manufacturer of the product, upstream suppliers, or vendors, or the LCA practitioner performing the study. Confidentiality issues are not relevant for life-cycle inventories conducted by companies using their own facility data for internal purposes. However, the use of proprietary data is a critical issue in inventories conducted for external use and whenever facility-specific data are obtained from external suppliers for internal studies. Consequently, current studies often contain insufficient source and documen- tation data to permit technically sound external review. Lack of technically sound data adversely affects the credibility of both the life-cycle inventories and the method for performing them. An individual company's trade secrets and competitive technologies must be protected. When collecting data (and later when reporting the results), the protection of confidential business infor- mation should be weighed against the need for a full and detailed analysis or disclosure of information. Some form of selective confidentiality agreements for entities performing life-cycle inventories, as well as formalization of peer review procedures, is often necessary for inventories that will be used in a public forum. Thus, industry data may need to undergo intermediate confi- dential review prior to becomingan aggregated data source for a document that is to be publicly released. Examples of private industry data sources include independent or internal reports, periodic measurements, accounting or engineering reports or data sets, specific measurements, and machine specifications. One particular issue of interest in considering industrial sources, whether or not a formal public data set is established, is the influence of industry and related technical associ- ations to enhance the accuracy, representativeness, and up-to-datedness of the collected data. Such associations may be willing, without providing specific data, to confirm that certain data (about which their members are knowledge- able) are realistic. 5.5 Public Industrial Data Technical books, reports, conference papers, and articles published in technical journals are a good source for information and data on industrial processes and SOURCING LIFE CYCLE INVENTORY DATA 113 activities. Most are publicly available, although data presented in these sources are often older, and they can be either too specific or not specific enough. Many of these documents give theoretical data rather than real data for processes. Such data may not be representative of actual processes or may deal with new technologies not commercially tested. In using the technical data sources in the following list, the analyst should consider the date, specificity, and relevancy of the data: • Kirk-Othme^s Encyclopedia of Chemical Technology • Periodical technical journals such as Journal of the Water Environment Federation • Proceedings from technical conferences • Textbooks on various applied sciences Frequently, the end user will not be able to supply specific information on inputs and outputs. However, the end user can provide data on user prac- tices from which inputs and outputs can be derived. Generally, the end user can be the source of related information from which the energy, materials, and pollutant release inventory can be derived. (An exception would be an institutional or commercial end user who may have some information on energy consumption or water effluents.) Market research firms can often provide qualitative and quantitative usage and customer preference data without the analyst having to perform independent market surveys (EPA 2006). 5.6 Dedicated LCI databases Since the early 1990s, LCA databases have proliferated in response to the grow- ing demand for life cycle information. These data sources have mostly ema- nated from Northeast Asia, North America, and Western Europe. In a global economy, of course, products and services are sourced from many countries. LCA databases mainly provide life cycle inventory (LCI) data, although char- acterization factors associated with life cycle impact assessment methods are often included as well. Tables 5.1 and 5.2 identify sources of LCI data, including public, as well as proprietary, or restricted-access, databases (updated from (Curran and Notten 2006)). Table 5.3 is based on the work by Dr. Joyce Coopers group at The University of Washington to identify publicly available information sources from databases, qualitative sources, and computational models that are suit- able for use in LCA (http://faculty.washington.edu/cooperjs/Definitions/ inventory squared.htm). Other providers of on-line lists of publicly available data sources (data- bases, qualitative sources, and computer models suitable for LCA) include the following: T ab le 5 .1 A va ila bl e na ti on al li fe c yc le in ve nt or y da ta ba se s (u pd at ed C ur ra n an d N ot te n 20 06 ). N am e A us tr al ia n L if e C yc le I nv en to ry D at a P ro je ct B U W A L 25 0 C an ad ia n R aw M at er ia ls D at ab as e ec oi nv en t E D IP G er m an N et w o rk on L if e C yc le In ve nt or y D at a Ja pa n N at io na l L C A P ro je ct W eb si te h tt p :/ /w w w .a u sl ci .c o m . a u / ht tp : / / s vi -V er pa ck un g. ch /d e/ S er v ic es / 1 & P ub li ka ti on en / ht tp : / / cr m d. uw at er lo o. ca / w w w .e co in ve nt .c h w w w .l ca -c en te r. dk w w w .l ci -n et w or k. de h tt p :/ /w w w .j em ai .o r. jp / en gl is h/ lc a/ pr oj ec t. cf m A va il ab il it y F re e F ee o r in cl ud ed w it h S im aP ro F re e w it h re gi st ra ti on L ic en se f ee L ic en se f ee O n- go in g F ee L an g u ag e E ng li sh G er m an , F re nc h E ng li sh , F re nc h E ng li sh D an is h G er m an , E ng li sh Ja pa ne se D at a F oc u s (i f an y) P ac ka gi ng m at er ia ls A lu m in u m , gl as s, p la st ic s, st ee l, an d w o o d N u m b er o f D at as et s 10 0 17 40 00 10 0 60 0 G eo g ra p h ic O ri g in A us tr al ia S w it ze rl an d C an ad a G lo b al / E u ro p e/ S w it ze rl an d D en m ar k G er m an y Ja pa n W n n r̂ w > en W C D on w H X > o od O O K or ea n L C I L C A Fo od 1 SP IN E@ C PM Sw is s A gr ic ul tu ra l L ife C yc le A ss es sm en t D at ab as e | (S A L C A ) T ha il N at io na l L C I D at ab as e U S LC I D at ab as e Pr oj ec t ht tp : / / w w w .k nc pc .r e. kr w w w .lc af oo d. dk ht tp : / / cp m da ta ba se .c pm . ch al m er s. se / ht tp :/ /w w w . ag ro sc op e. ad m in .c h/ oe ko bi la nz en / ht tp :/ /w w w . th ai lc id at ab as e. ne t/ w w w .n re l.g ov /l ci O n- go in g Fr ee Fr ee A va ila bl e th ro ug h ec oi nv en t o r w it h pr oj ec t co op er at io n Fr ee w it h co nt ac t K or ea n, E ng lis h D an is h, E ng lis h E ng lis h G er m an , E ng lis h, Fr en ch , It al ia n T ha i, E ng lis h E ng lis h E ne rg y, ch em ic al s, m et al , p ap er , ru bb er , po ly m er s, el ec tr on ic / el ec tr ic , co ns tr uc ti on , pr od uc ti on pr oc es s, de liv er y, di sp os al , a nd ut il it y Fo od p ro du ct s an d pr oc es se s - A gr ic ul tu re 15 8 70 0 70 0 30 0 K or ea D en m ar k G lo ba l Sw it ze rl an d T ha il an d U SA o s o •n w n o r1 w < w H O S I O i T ab le 5 .2 I nd us tr y or ga ni za ti on s' d at ab as es . In d u st ry O rg an iz at io n A m er ic an P la st ic s C ou nc il ( A P C ) E P D -N or w ay E ur op ea n A lu m in iu m A ss oc ia ti on ( E A A ) E ur op ea n C op pe r 1 In st it ut e (E C I) E ur op ea n F ed er at io n of C or ru ga te d B oa rd M an uf ac tu re rs (F E FC O ) In te rn at io na l Ir on a nd S te el I ns ti tu te ( II SI ) In te rn at io na l Z in c A ss oc ia ti on W eb si te A va il ab le f ro m U S L C I D at ab as e (n re l. go v/ lc i) w w w .e pd -n or ge .n o w w w .a lu m in iu m .o rg w w w .c op pe r- li fe -c yc le .o rg w w w .f ef co .o rg w w w .w or ld st ee l. or g in fo @ iz a. co m A va il ab il it y F re e F re e F re e F re e w it h co nt ac t F re e F re e w it h co nt ac t A va il ab le t o L C A pr ac ti ti on er s on r eq ue st L an gu ag e E ng li sh N or w eg ia n, E ng li sh E ng li sh E ng li sh E ng li sh E ng li sh E ng li sh P ro du ct G ro u p or S ec to r P ol ym er s N o rw eg ia n bu si ne ss (s ev er al s ec to rs ) A lu m in iu m p ro du ct io n C o p p er t ub es , sh ee ts an d w ir e C or ru ga te d B oa rd S te el Z in c G eo gr ap h ic C ov er ag e A m er ic a N o rw ay an d E ur op e E ur op e E ur op e E ur op e G lo ba l G lo ba l O N w n n w > C D C D w C D C D w H X > a « o o IS SF I nt er na ti on al St ai nl es s st ee l Fo ru m (I SS F) K C L (E co D at a) N ic ke l I ns ti tu te Pl as tic sE ur op e (f or m er ly A PM E ) V ol vo E PD s W or ld S te el C ar bo n Fo ot pr in t w w w .w or ld st ai nl es s. or g/ ht tp : / / w w w .k cl .f i/ ht tp : / / w w w .n ic ke lin st i tu te .o rg w w w .p la st ic se ur op e. or g ht tp :/ /w w w . vo lv ot ru ck s. co m / de al er s- vt c / e n- gb / V T B C -E as tA ng lia / ab ou tu s/ en vi ro nm en t/ en vi ro nm en ta l_ pr od uc t_ de cl ar at io n ht tp :/ /w w w . w or ld au to st ee l. or g/ E nv ir on m en t/ L if e- C yc le -A ss es sm en t/ w or ld st ee l- re le as es - da ta se ts -t o- he lp -l ow er - ca rb on -f oo tp ri nt .a sp x Fr ee w it h co nt ac t Fe e Fr ee w it h co nt ac t Fr ee Fr ee Fr ee b y re qu es t E ng lis h, C hi ne se , Ja pa ne se E ng lis h E ng lis h E ng lis h E ng lis h E ng lis h St ai nl es s st ee l P ul p an d pa pe r N ic ke l Pl as tic s T ru ck s an d bu ss es St ee l p ro du ct s G lo ba l 1 F in ni sh / N or di c G lo ba l E ur op e E ur op e G lo ba l o s I r1 +* \ W n 8 r1 w w H O 3 I 118 LIFE CYCLE ASSESSMENT HANDBOOK - The European Commission's Joint Research Center developed the European Reference Life Cycle Database (ELCD), which is com- prised of LCI data for key materials, energy carriers, transport, and waste management. The respective data sets are officially provided and approved by the named industry association. New data continue to be added. The data sets are accessible free of charge and without access or use restrictions for all LCA practitio- ners. The ELCD core database version II can be found at: http:// lca.jrc.ec.europa.eu/lcainfohub/datasetArea.vm. - The UNEP/SETAC Database Registry aims to (1) aid users world- wide in finding the most suitable (and therefore best quality) data and (2) assist data providers worldwide in finding users (http:// lca-data.org). 5.7 Using Non-LCI Data in LCAs Government documents and databases provide data on broad categories of processes and are publicly available. Most government documents are published on a periodic basis, e.g., annually, biennially, or every four years. However, the data published within them tend to be at least several years old. Furthermore, the data found in these documents may be less specific and less accurate than industry data for specific facilities or groups of facilities. However, depending on the purpose of the study and the specific data objec- tives, these limitations may not be critical. All studies should note the age of the data used. Government databases include both non-bibliographic types where the data items themselves are contained in the database and bibliographic types that consist of references where data may be found. In a study conducted for the US Environmental Protection Agency, Boguski identi- fied twelve data sources that have utility in developing LCA data sets for the US (Boguski 2000)2. The data sources Boguski evaluated included the following: • Aerometric Information Retrieval System (AIRS) - US Environ- mental Protection Agency • Permit Compliance System (PCS) - US Environmental Protection Agency • Biennial Reporting System (BRS) - US Environmental Protection Agency • Toxics Release Inventory (TRI) Database - US Environmental Protection Agency 2 Other useful U.S. government sources include: U.S. Department of Commerce's Census of Manufacturers, U.S. Bureau of Mines' Census of Mineral Industries, and the U.S. Department of Energy's Monthly Energy Review. SOURCING LIFE CYCLE INVENTORY DATA 119 • Industrial Assessment Center Database (IAC) - US Department of Energy • Manufacturing Energy Consumption Survey (MECS) - US Energy Information Administration • Reasonably Available Control Technology / Best Available Control Technology/Lowest Achievable Emissions Rate (RACT/BACT/ LAER) Clearinghouse (RBLC) - US Environmental Protection Agency • Compilation of Air Pollutant Emission Factors AP-42, Volume I: Stationary Point and Area Sources - US Environmental Protection Agency • Compilation of Air Pollutant Emission Factors AP-42, Volume II: Mobile Sources - US Environmental Protection Agency • Locating and Estimating Air Emissions from Sources (A series of L&E documents) - US Environmental Protection Agency • Factor Information Retrieval (FIRE) - US Environmental Protection Agency • Sector Notebooks - US Environmental Protection Agency Boguski demonstrated that it is possible to extract meaningful information from public databases for use in LCA studies. It is even possible to develop LCI data sets for some products by using information from public databases. Public databases have several advantages. They are accessible to anyone who wishes to check LCI results. They typically include many more of the specific emissions from industrial facilities than are included in most private LCI databases. They include data directly from U.S. facilities. There is no need to try to convert European data to U.S. conditions. However, there are disadvantages to using public databases. The organization and presentation of data in public databases often makes it difficult to express the values from the various databases in terms that are generally useful for LCA. One chal- lenge is being able to link energy and emission values to production. For example, the MECS database reports annual energy use for industry groups. Likewise, AIRS, TRI, and BRS report annual emissions. The PCS database reports monthly monitoring values, which may be averaged to obtain annual emission estimates. None of these databases ties energy use or emissions to production. Production information is difficult to obtain. Production on a facility level is usually considered confidential information and is not usually published. The United States Census Bureau reports production, in mass units by SIC code, for only a few industry groups. In addition, facilities are not reported using unique identifiers, leading to difficulty in linking data sources with produc- tion rates (for example, when a facility is sold it is reported multiple times under different ownership names). A method for grouping facility data into logical industry groupings and linking the grouped data to grouped produc- tion values would benefit the LCA community and still provide confidentiality to industry. T ab le 5 .3 L if e cy cl e i n ve nt or y da ta s ou rc es , p re pa re d by J oy ce C oo pe r, U ni ve rs it y of W as hi ng to n. A va il ab le a t ht tp :/ /f ac ul ty .w as hm gt on .e du /c oo pe rj s/ D ef in it io ns /i nv en to ry sq ua re d. ht m ( ac ce ss ed J an ua ry 2 01 2) . D at ab as e T it le A er om et ri c In fo rm at io n R et ri ev al S ys te m (U S E PA ) A lt er na ti ve F ue l & A dv an ce d V eh ic le s D at a C en te r (U S D O E ) A nn ua l E ne rg y O ut lo ok ( U S D O E E ne rg y In fo rm at io n A dm in is tr at io n) B ienn ia l R ep or ti ng S ys te m ( U S E PA ) P ro ce ss es , F ac ili ty , Se ct or , o r M at er ia l- B as ed F ac il it y M at er ia ls P ro ce ss a nd S ec to r F ac il it ie s A p p li ca b il it y M at er ia ls A cq u. & P ro ce ss , C on st ru ct & M an uf ., E ne rg y, T ra ns po rt E ne rg y, T ra ns po rt E ne rg y, T ra ns po rt E nd -o f- L if e C at eg or ie s of S ec to rs , F ac il it ie s, or U n it P ro ce ss es In du st ri al , ag ri cu lt ur al , an d tr an sp or ta ti on so ur ce s V eh ic le F ue ls In du st ri al a nd ot he r en er gy us e T re at m en t, S to ra ge , an d D is po sa l fa ci li ti es Sp ec if ic S ec to rs , F ac il it ie s, U n it P ro ce ss or o r M at er ia ls C ap tu re d E xa m pl e of f ac il it ie s (s ou rc es ) in cl ud e el ec tr ic po w er p la nt s, s te el m il ls , f ac to ri es , an d un iv er si ti es B io di es el , E le ct ri c F ue l, E th an ol . H yd ro ge n, M et ha no l, N at ur al G as ( C N G /L N G ), P ro pa ne (L P G ), P -S er ie s, S ol ar F ue l R es id en ti al . C om m er ci al , In du st ri al , T ra ns po rt at io n, D el iv er ed E ne rg y C on su m pt io n fo r al l se ct or s, E le ct ri c G en er at or s. In p u ts a n d O u tp u ts In cl u d ed C ap tu re s O u tp u t P ro du ct s, C ap tu re s O u tp u t W as te s (A ir em is si on f ac to rs pr es en te d pe r un it o f ou tp ut ) C ap tu re s in pu t E ne rg y, C ap tu re s In pu t M at er ia ls o r la nd U se , C ap tu re s O u tp u t P ro du ct s, C ap tu re s O u tp u t W as te s (A ir em is si on f ac to rs pr es en te d pe r un it o f ou tp ut ) C ap tu re s In pu t E ne rg y, C ap tu re s O u tp u t P ro du ct s (e ne rg y) , C ap tu re s O u tp u t W as te s C ap tu re s O u tp u t W as te s ho o w n n r w > C D Q /5 W w H > a O o M ea t P ac ki ng , F lu id M il k, C an n ed F ru it a n d ve ge ta b le s, f ru it a n d ve ge ta b le j ui ce s, w et co rn m il li n g, b re ad b ak in g, c ak es a n d p ie s b ak in g, c an e su ga r re fi ni ng , b ee t su ga r re fi ni ng , so yb ea n oi l m il ls , M al t B ev er ag es , w ea vi n g m il ls , fi ni sh in g m il ls , lo gg in g ca m p s, s aw m il ls a n d p la n n in g m il ls , w oo d p ro d u ct s, N E C ( F ib er bo ar d) , P u lp M il ls , pa pe r B oa rd M il ls , C or ru ga te d, S ol id F ib er B ox es , B u il d in g P ap er , A lk al ie s an d C ho ri ne , In or ga ni c G as es , I no rg an ic P ig m en ts , In du st ri al in or ga ni c, c h em ic al s, P la st ic M at er ia ls a n d R es in s, S yn th et ic R ub be rs , C el lu lo si c M an - m ad e F ib er s, o rg an ic F ib er s, P ha rm ac eu ti ca l P re pa ra ti on s, C yc li c C ru d es a n d In te rm ed ia te s, In du st ri al O rg an ic C h em ic al s, F er ti liz er s, C he m ic al P re pa ra ti on s, P et ro le u m R ef in in g, P av in g M at er ia ls a n d B lo ck s, T ir es a n d In ne r T ub es , F ab ri ca te d R ub be r P ro du ct s, M is ce ll an eo u s P la st ic P ro d u ct s, G la ss , C em en t, B ri ck a n d St ru ct ur al C la y T ile , L im e, G yp su m P ro du ct s, M in er al W oo l, B la st F ur na ce s an d St ee l M il ls , E le ct ro m et al lu rg ic al P ro du ct s, G ra y Ir on F ou n d ri es , P ri m ar y C op pe r, P ri m ar y A lu m in u m , S ec on d ar y N on -F er ro u s M et al s, A lu m in u m F in is h F or m in g, I ro n an d St ee l F or gi ng , F ar m M ac hi ne ry a n d E qu ip m en t, C on st ru ct io n M ac hi ne ry , M ot or V eh ic le s an d C ar B od ie s, M ot or V eh ic le s P ar ts a n d A cc es s, P ho to gr ap hi c F il m , P h ot og ra p h ic E qu ip m en t T ab le 5 .3 (c on t.) L ife c yc le in ve nt or y da ta s ou rc es , p re pa re d by J oy ce C oo pe r, U ni ve rs ity o f W as hi ng to n. A va ila bl e at ht tp :/ /f ac ul ty .w as hi ng to n. ed u/ co op er js /D ef in it io ns /i nv en to ry sq ua re d. ht m ( ac ce ss ed J an ua ry 2 01 2) . D at ab as e T it le E -G R ID : E m is si on s & G en er at io n R es ou rc e In te gr at ed D at ab as e (U S E PA ) ID E M A T (T U D el ft , T he N et he rl an ds ) N at io na l W at er U se I nf or m at io n P ro gr am ( U SG S W at er R es ou rc es D iv is io n) T ox ic s R el ea se In ve nt or y (U S E PA ) P ro ce ss es , F ac ili ty , Se ct or , o r M at er ia l- B as ed P ro ce ss a nd F ac il it y M at er ia ls S ec to r F ac il it y A p p li ca b il it y E ne rg y M at er ia ls A cq u. & P ro ce ss , C on st ru ct . & M an uf ., E nd -o f- L if e M at er ia ls A cq u. & P ro ce ss , A gr ic ul tu re , C on st ru ct . & M an uf ., U se , E ne rg y C on st ru ct . & M an uf . C at eg or ie s of S ec to rs , F ac il it ie s, or U n it P ro ce ss es E ne rg y P ro du ct io n In du st ri al , ag ri cu lt ur al , co m m er ci al , re si de nt ia l, ut il it ie s M an uf ac tu ri ng an d go ve rn - m en t fa ci li ti es S p ec if ic S ec to rs , F ac il it ie s, U n it P ro ce ss or o r M at er ia ls C ap tu re d A ll U S so ur ce s - ai r em is si on s an d kW p ro du ct io n by t ec hn ol og y ty pe In p u ts a n d O u tp u ts In cl u d ed C ap tu re s O u tp u t P ro du ct s, C ap tu re s O u tp u t W as te s (A ir em is si on s) P ro vi de s P ro ce ss F lo w D ia gr am s, C ap tu re s In pu t M at er ia ls o r L an d U se C ap tu re s In pu t M at er ia ls or l an d U se ( su rf ac e an d g ro u n d w at er us e) , C ap tu re s O u tp u t W as te s (r ec yc li ng ) C ap tu re s In pu t M at er ia ls or l an d U se , C ap tu re s O u tp u t P ro du ct s (a nn ua l ra ti o) , C ap tu re s O u tp u t W as te s ro W n n r· w > Q /i W C D w H > Ö od O O U S E co no m ic C en su s (U S C en su s B ur ea u) Fa ci lit y M at er ia ls A cq u. & P ro ce ss , C on st ru ct . & M an uf . M in in g, co ns tr uc tio n an d m an uf ac tu ri ng T he M in in g se ct or o f th e 19 97 E co no m ic C en su s co ve rs a ll m in in g es ta bl is hm en ts o f co m pa ni es w ith o ne o r m or e pa id e m pl oy ee s. M in in g is de fin ed a s th e ex tr ac tio n of n at ur al ly o cc ur - ri ng m in er al s ol id s, s uc h as c oa l a nd o re s; liq ui d m in er al s, s uc h as p et ro le um , a nd g as es, su ch a s na tu ra l g as . T he te rm m in in g is u se d in th e br oa d se ns e to in cl ud e qu ar ry in g, w el l o pe ra tio ns , b en ef ic ia tin g (i. e. , c ru sh in g, sc re en in g, w as hi ng , a nd f lo at at io n) , a nd o th er pr ep ar at io ns c us to m ar ily p er fo rm ed a t t he m in e si te o r a s pa rt o f th e m in in g ac tiv iti es ( ht tp :/ / w w w .c on su s. go v/ pr od /w w w /a bs /9 7/ ec m in . ht m l) . T he c on st ru ct io n re po rt s be lo w in cl ud e ne w c on st ru ct io n w or k, a dd iti on s, a lte ra - tio ns , a nd r ep ai rs . E st ab lis hm en ts i de nt if ie d as c on st ru ct io n m an ag em en t f irm s ar e al so in cl ud ed . T he c on st ru ct io n se ct or is d iv id ed in to th re e ty pe s of a ct iv ity o r su bs ec to rs . T he su bs ec to rs a re th e B ui ld in g, D ev el op in g, a nd G en er al C on st ru ct in g, H ea vy C on st ru ct io n, a nd Sp ec ia l T ra de C on tr ac to rs . T he a re a re po rt s fo r th e co ns tr uc tio n in du st ri al c on ta in s ta te a nd re gi on al le ve l d at a (h tt p: // ce ns us .g ov /p ro d/ w w w /a bs /9 7/ ec m an i. ht m l) . E st ab lis hm en ts in th e m an uf ac tu ri ng s ec to r ar e of te nd es cr ib ed as p la nt s, fa ct or ie s, o r m ill s an d ty pi ca lly u se po w er -d ri ve n m ac hi ne s an d m at er ia ls -h an dl in g eq ui pm en t. A ls o in cl ud ed in th e m an uf ac tu ri ng se ct or a re s om e es ta bl is hm en ts th at m ak e pr od - uc ts b y ha nd , l ik e cu st om ta ilo rs a nd th e m ak er s of c us to m d ra pe ri es ; s om e es ta bl is hm en ts li ke ba ke ri es a nd c an dy s to re s th at m ak e pr od uc ts on th e pr em is es m ay b e in cl ud ed ( ht tp :/ /w w w . ce ns us .g ov /p ro d/ w w w /a bs /9 7/ ec m an i. ht m l) C ap tu re s in pu t E ne rg y, C ap tu re s In pu t M at er ia ls o r la nd U se , C ap tu re s O ut pu t Pr od uc ts O s o r· w n n r1 w w H O I CO (C on tin ue d) T ab le 5 .3 (c on t.) L ife c yc le in ve nt or y da ta s ou rc es , p re pa re d by J oy ce C oo pe r, U ni ve rs ity o f W as hi ng to n. A va ila bl e at ht tp :/ /f ac ul ty .w as hi ng to n. ed u/ co op er js /D ef in it io ns /i nv en to ry sq ua re d. ht m ( ac ce ss ed J an ua ry 2 01 2) . D at ab as e T it le A nn ua l an d M on th ly E ne rg y R ev ie w s (U S D O E E ne rg y In fo rm at io n A dm in is tr at io n) A nn ua l E st im at es of G lo ba l A nt hr op og en ic M et ha ne E m is si on s (C en te r fo r E ne rg y an d E nv ir on m en ta l S tu di es , B os to n | U ni v. ) E M E P /C O R IN A IR A tm os ph er ic E m is si on F ac to rs fo r E ur op e (E E A ) P ro ce ss es , F ac ili ty , Se ct or , o r M at er ia l- B as ed P ro ce ss a nd S ec to r S ec to r P ro ce ss A p p li ca b il it y E ne rg y, T ra ns po rt A gr ic ul tu re , E nd -o f- L if e, E ne rg y C on st ru ct . & M an uf . C at eg or ie s of S ec to rs , F ac il it ie s, or U n it P ro ce ss es In du st ri al a nd ot he r en er gy us e E ne rg y, a gr i- cu lt ur al , an d w as te s to ra ge In du st ri al s ou rc es S p ec if ic S ec to rs , F ac il it ie s, U n it P ro ce ss or o r M at er ia ls C ap tu re d E ne rg y P ro du ct io n by S ou rc e, E ne rg y C on su m pt io n by s ou rc e, E ne rg y C on su m pt io n by E nd -U se S ec to r: h ou se ho ld a n d m ot or ve hi cl es , E ne rg y R es ou rc es : cr ud e O il a nd N at ur al G as F ie ld C ou nt s; P ro du ct io n, R es er ve s, R ec ov er y of O il a n d G as , L iq ui d an d G as eo us h yd ro ca rb on , u ra n iu m , an d pe tr ol eu m F la ri ng a nd v en ti ng o f N at ur al G as , O il a nd ga s S up pl y S ys te m s, E xc lu di ng F la ri ng , C oa l M in in g, B io m as s B ur ni ng , li ve st oc k F ar m in g, R ic e F ar m in g an d R el at ed A ct iv it ie s, L an df il ls In du st ri al c om bu st io n pl an t an d pr oc es se s w it h co m bu st io n; c om bu st io n in b oi le rs , g as ; tu rb in es an d st at io na ry e ng in es ; c om bu st io n pl an ts 9 30 0 M W ; c om bu st io n pl an ts 9 5 0 M W a nd < 30 0 M W ; c om bu st io n pl an ts < 50 M W G as t ur bi ne s; st at io na ry e ng in es ; P ro ce ss f ur na ce s w it ho ut co nt ac t (1 ); R ef in er y pr oc es se s fu rn ac es , co ke ov en f ur na ce s; b la st f ur na ce s co w pe rs ; P la st er fu rn ac es , P ro ce ss es w it h co nt ac t (2 ); S in te r pl an t; R eh ea ti ng f ur na ce s st ee l an d ir on ; G ra y ir on fo un dr ie s; p ri m ar y le ad p ro du ct io n, P ri m ar y zi nc p ro du ct io n, P ri m ar y co op er p ro du ct io n, In p u ts a n d O u tp u ts In cl u d ed C ap tu re s In pu t E ne rg y, C ap tu re s O u tp u t P ro du ct s, C ap tu re s O u tp u t W as te s C ap tu re s O u tp u t W as te s (m et ha ne t o ai r) « n n w > Q D W w H X > a o o C hi ef C le ar in gh ou se fo r In ve nt or ie s an d E m is si on F ac to rs (U S E PA ) D at ab as e of U S G re en ho us e G as E m is si on s (U S D O E E ne rg y In fo rm at io n A dm in is tr at io n) L if e C yc le M an ag em en t O f M un ic ip al S ol id W as te P ro ce ss a nd F ac il it y S ec to r an d M at er ia ls P ro ce ss M at er ia ls A cq u. & P ro ce ss , A gr ic ul tu re , C on st ru ct . & M an uf ., U se , E nd -o f- L if e, E ne rg y, T ra ns po rt M at er ia ls A cq u. & P ro ce ss , C on st ru ct . & M an uf ., U se , E ne rg y, T ra ns po rt E nd -o f- L if e In du st ri al , ag ri - cu lt ur al a nd tr an sp or ta ti on so ur ce s In du st ri al , co m m er ci al , re si de nt ia l, tr an sp or ta ti on , an d ut il it ie s M un ic ip al so li d w as te m an ag em en t S ec on da ry l ea d pr od uc ti on ; S ec on da ry z in c pr od uc ti on ; se co nd ar y co op er p ro du ct io n, se co nd ar y al um in um p ro du ct io n, c em en t, L im e (i nc lu di ng i ro n an d st ee l an d pa pe r p u lp in du st ri es ); A sp ha lt c on cr et e pl an ts ; R at g la ss ; M in er al w oo l (e xc ep t bi nd in g) ; B ri ck s an d ti le s, F in e ce ra m ic m at er ia l; P ap er m il l in du st ry (d ry in g pr oc es s) ; a lu m in a pr od uc ti on P et ro le um , co al , g eo th er m al , na tu ra l ga s, c em en t pr od uc ti on, na tu ra l ga s, g as f la ri ng , ke ro se ne , je t fu el s, t ra ns po rt at io n (h ig hw ay v eh ic le s, a ir tr an sp or t, v es se ls ), f or es t fi re s, L P G , a nd m u ch m or e C ol le ct io n; M at er ia ls R ec ov er y F ac il it y; E ne rg y M od el ; T ra ns po rt at io n; T ra ns fe r S ta ti on s; C om bu st io n; C om po st ; R em an uf ac tu ri ng C ap tu re s O u tp u t P ro du ct s, C ap tu re s O u tp u t W as te s (A ir em is si on f ac to rs pr es en te d pe r un it o f ou tp ut ) C ap tu re s O u tp u t W as te s (g re en ho us e ga se s) P ro vi de s a D es cr ip ti on of U ni t P ro ce ss es , P ro vi de s P ro ce ss F lo w D ia gr am s, C ap tu re s In pu t E ne rg y, C ap tu re s In pu t M at er ia ls f or l an d U se , C ap tu re s O u tp u t P ro du ct s; C ap tu re s O u tp u t W as te s (C on tin ue d) o 8 o r W n n r1 w I— I < W H O 3 ! CJ 1 T ab le 5 .3 (c on t.) L ife c yc le in ve nt or y da ta s ou rc es , p re pa re d by J oy ce C oo pe r, U ni ve rs it y of W as hi ng to n. A va ila bl e at ht tp :/ /f ac ul ty .w as hi ng to n. ed u/ co op er js /D ef in it io ns /i nv en to ry sq ua re d. ht m ( ac ce ss ed J an ua ry 2 01 2) . D at ab as e T it le N A E L T h eU K E m is si on F ac to rs D at ab as e (A E A T ec hn ol og y) N at io na l A gr ic ul tu ra l S ta ti st ic s S er vi ce : A gr ic ul tu ra l C he m ic al U se D at ab as e (U SD A ) P ro ce ss es , F ac ili ty , Se ct or , o r M at er ia l- B as ed P ro ce ss M at er ia ls A p p li ca b il it y M at er ia ls A cq u. & P ro ce ss , A gr ic ul tu re , C on st ru ct . & M an uf ., U se , E nd -o f- L if e, E ne rg y, T ra ns po rt A gr ic ul tu re C at eg or ie s of S ec to rs , F ac il it ie s, or U n it P ro ce ss es In du st ri al , ag ri cu lt ur al , tr an sp or ta ti on , an d se rv ic es (h os pi ta ls ) so ur ce s A gr ic ul tu re S p ec if ic S ec to rs , F ac il it ie s, U n it P ro ce ss or o r M at er ia ls C ap tu re d M ob il e S ou rc es : r oa d T ra ff ic , 'C ol d S ta rt ' em is - si on s, ' ho t so ak ' em is si on s, r ai l tr af fi c, a ir po rt s, sh ip s. A re a so ur ce s: E m is si on s fr om l ar ge n u m b er s th at a re o f lo w s ig ni fi ca nc e of s m al l em it te rs ( i.e ., do m es ti c ga s bo il er s) o r fr om ot he r id en ti fi ab le a re as ( i.e ., fa rm la nd o r la nd fi ll s it es ) ar e ag gl om er at ed t og et he r, b y ty pe , b as ed o n w hi ch n at io na l gr id s qu ar e th ey fa ll in . P oi nt s ou rc es : M an y of t he e m is si on s to t he a tm os ph er e re su lt in g fr om i nd us tr ia l pr oc es se s an d th e co m bu st io n of f os si l fu el s ar e no t un if or m ly s pr ea d ac ro ss u rb an a re as bu t co nc en tr at ed a t pa rt ic ul ar p oi nt s. T he se po in t so ur ce s in cl ud e ce nt ra l he at in g pl an ts se rv in g la rg e g ro u p s of b ui ld in gs , s uc h as ho sp it al s, a n d bo il er p la nt s su pp ly in g pr oc es s he at t o in du st ry . T he y al so i nc lu de i nd us tr ia l pr oc es se s w hi ch r eq ui re a ut ho ri za ti on u n d er th e E nv ir on m en ta l P ro te ct io n A ct 1 99 0. B y cr op t yp e an d by p es ti ci de In p u ts a n d O u tp u ts In cl u d ed C ap tu re s O u tp u t P ro du ct s, C ap tu re s O u tp u t W as te s (a ir em is si on f ac to rs pr es en te d pe r un it o f ou tp ut ). C ap tu re s In pu t m at er ia ls or l an d U se ( pe st ic id e us e) O N w n n r· w > C D W CT > C D w H > a a) O O C en su s of A gr ic ul tu re (U SD A ) M on th a n d St at e C ur re nt E m is si on s T re nd s (U S D O E A rg on n e N at io n al L ab or at or y) N at io n al C oa st al P ol lu ta nt D is ch ar ge In ve nt or y (N O A A ) N at ur al R es ou rc es In ve nt or y (U SD A ) F ac ili ty a nd Se ct or P ro ce ss es F ac ili ty Se ct or M at er ia ls A cq u . & P ro ce ss , A gr ic ul tu re , C on st ru ct . & M an uf . U se , E nd -o f- L if e, E ne rg y, T ra ns po rt C on st ru ct . & M an u f. , E ne rg y, T ra ns po rt M at er ia ls A cq u . & P ro ce ss , A gr ic ul tu re , C on st ru ct . & M an uf ., U se M at er ia ls A cq u . & P ro ce ss , A gr ic u lt u re , C on st ru ct . & M an uf ., u se , E nd -o f- L if e, E ne rg y, T ra ns po rt A gr ic ul tu re , pa st ur e, f or es t, tr an sp or ta ti on , d ef en se a n d in du st ri al a re as E le ct ri c u ti li ti es , in du st ri al f ue l co m b u st io n , co m m er ci al / re si de nt ia l fu el co m b u st io n , In du st ri al p ro - ce ss es , t ra ns - po rt at io n, a n d m is ce ll an eo u s In du st ri al a n d ag ri cu lt ur al em is si on so u rc es F ed er al a n d n on - F ed er al l an d s in cl u d in g ag ri cu lt ur e, pa st ur e, f or es t, et c. C ap tu re s in p u t M at er ia ls or L an d u se ( la nd u se ) C ap tu re s O u tp u t W as te s (n it ro ge n ox id es , su lf ur o xi d es , an d n on m et h an e V O C s) C ap tu re O u tp u t W as te s (w at er e m is si on s) C ap tu re s In pu t M at er ia ls or l an d U se (C on tin ue d) O cj B O r· W n n r1 w I— I < w S o I T ab le 5 .3 (c on t.) L ife c yc le in ve nt or y da ta s ou rc es , p re pa re d by J oy ce C oo pe r, U ni ve rs ity o f W as hi ng to n. A va ila bl e at ht tp :/ /f ac ul ty .w as hi ng to n. ed u/ co op er js /D ef in it io ns /i nv en to ry sq ua re d. ht m ( ac ce ss ed J an ua ry 2 01 2) . D at ab as e T it le R ec yc li ng P ro ce ss es (R ec yc li ng D at a M an ag em en t C or po ra ti on ) R eP IS : R en ew ab le E le ct ri c P la nt In fo rm at io n S ys te m ( N at io na l R en ew ab le E ne rg y L ab or at or y) R en ew ab le R es ou rc e D at a C en te r (N at io na l R en ew ab le E ne rg y L ab or at or y) E ng in ee re d M at er ia ls A bs tr ac ts P ro ce ss es , F ac ili ty , Se ct or , o r M at er ia l- B as ed M at er ia ls P ro ce ss es a nd F ac il it ie s S ec to r P ro ce ss es a nd M at er ia ls A p p li ca b il it y E nd -o f- L if e E ne rg y E ne rg y M at er ia ls A cq u. & P ro ce ss , A gr ic ul tu re , co ns tr uc t. & M an uf ., us e, E nd -o f- L if e, E ne rg y, T ra ns po rt C at eg or ie s of S ec to rs , F ac il it ie s,or U n it P ro ce ss es In du st ri al a nd ot he r m at er ia ls E le ct ri c pl an ts E ne rg y ef fi ci en cy an d re ne w ab le en er gy In du st ry , ut il it ie s, ag ri cu lt ur e, se rv ic es , e tc . S p ec if ic S ec to rs , F ac il it ie s, U n it P ro ce ss or o r M at er ia ls C ap tu re d G eo th er m al P la nt s, H y d ro P la nt s, L an df il l M et ha ne P la nt s, P ho to vo lt ai c P la nt s, S ol ar T he rm al P la nt s, w as te t o E ne rg y P la nt s, W in d P la nt s, W oo d an d A g W as te P la nt s In p u ts a n d O u tp u ts In cl u d ed C ap tu re s In pu t M at er ia ls or L an d U se , C ap tu re s O u tp u t P ro du ct s C ap tu re s O u tp u t P ro du ct s (e ne rg y) P ro vi de s a D es cr ip ti on of U ni t P ro ce ss es , C ap tu re s In pu t E ne rg y, C ap tu re s In pu t M at er ia ls o r L an d U se , C ap tu re s O u tp u t P ro du ct s, C ap tu re s O u tp u t W as te s P ro vi de s a D es cr ip ti on of U ni t pr oc es se s, C ap tu re s In pu t E ne rg y, C ap tu re s In pu t M at er ia ls o r la nd U se , C ap tu re s O u tp u t P ro du ct s, C ap tu re s O u tp u t W as te s 0 0 w n >< n r1 w > on W </> to w H X > o o o K ir k- O th m er E nc yc lo pe di a of C he m ic al T ec hn ol og y L on g T er m W or ld O il S u p p ly ( A R es ou rc e B as e/ P ro du ct io n P at h A n al ys is ( U S D O E E ne rg y In fo rm at io n A dm in is tr at io n) P ro ce ss , M at er ia ls P ro ce ss M at er ia ls A cq u . & P ro ce ss , C on st ru ct . & M an uf ., E nd -o f- L if e, E ne rg y, T ra ns po rt E ne rg y C h em ic al p ro ce ss es O il P ro d u ct io n A re as o f ch em ic al t ec h n ol og y th at w il l d ea l w it h 1 in du st ri al p ro d u ct s, n at ur al m at er ia ls , a n d p ro - ce ss es i n su ch f ie ld s as : a gr ic ul tu ra l ch em ic al s, ch em ic al e n gi n ee ri n g, c oa ti n gs a n d in k s, c om - p os it e m at er ia ls , c os m et ic a n d p h ar m ac eu ti ca ls , d ye s, p ig m en ts a n d br ig ht en er s, e co lo gy a n d in du st ri al h yg ie n e, e n er gy c on se rv at io n an d te ch n ol og y, f at s an d w ax es , f er m en ta ti on a nd en zy m e te ch n ol og y, f ib er s, t ex ti le s an d le at he r, 1 fo od a n d an im al n ut ri ti on , f os si l fu el s an d 1 d er iv at iv es , g la ss , c er am ic s an d ce m en t, i n d u s- 1 tr ia l i no rg an ic c h em ic al s, i nd us tr ia l or ga ni c 1 ch em ic al s, m et al s, m et al lu rg y an d m et al a ll oy s, 1 p la st ic s an d el as to m er s, s em ic on d u ct or s an d 1 em u ls io n te ch n ol og y, w at er s u p p ly , pu ri fi ca - ti on a n d re u se , w oo d , pa pe r, a n d in du st ri al 1 ca rb oh yd ra te s. A ls o in cl u d es m is ce ll an eo u s 1 to pi cs : in st ru m en ta ti on a n d qu al it y co nt ro l, in fo rm at io n re tr ie va l, m ai n te n an ce , m ar ke t re se ar ch , m at er ia l al lo ca ti on a n d su p p ly , le ga l is su es , p ro ce ss d ev el op m en t an d d es ig n , 1 p ro d u ct d ev el op m en t an d te ch ni ca l se rv ic e, 1 re se ar ch a n d op er at io n s m an ag em en t (s ys te m s 1 m an ag em en t, n et w or k s, e tc .) , a n d tr an sp or ta - 1 ti on o f ch em ic al p ro d u ct s. P ro vi d es a D es cr ip ti on of U n it P ro ce ss es , P ro vi d es P ro ce ss F lo w D ia gr am s, C ap tu re s In pu t E ne rg y, C ap tu re s In pu t M at er ia ls o r L an d U se , C ap tu re s O u tp u t P ro d u ct s, C ap tu re s O u tp u t W as te s P ro vi d es a D es cr ip ti on of U n it P ro ce ss es , C ap tu re s O u tp u t P ro du ct s (o il) (C on tin ue d) T ab le 5 .3 (c on t.) L ife c yc le in ve nt or y da ta s ou rc es , p re pa re d by J oy ce C oo pe r, U ni ve rs ity o f W as hi ng to n. A va ila bl e at ht tp :/ /f ac ul ty .w as hi ng to n. ed u/ co op er js /D ef in it io ns /i nv en to ry sq ua re d. ht m ( ac ce ss ed J an ua ry 2 01 2) . D at ab as e T it le M in er al I nd us tr y S ur ve y (U SG S) M un ic ip al S ol id W as te F ac tb oo k (U S E PA ) M un ic ip al S ol id W as te S ur ve y (U S E PA O ff ic e of S ol id W as te ) N et G en er at io n an d U ti li ty R et ai l S al es (U S C en su s B ur ea u) S ea rc h U S P at en ts P ro ce ss es , F ac ili ty , Se ct or , o r M at er ia l- B as ed M at er ia ls F ac il it ie s M at er ia ls P ro ce ss P ro ce ss es a nd M at er ia ls A p p li ca b il it y M at er ia ls A cq u. & P ro ce ss E nd -o f- L if e E nd -o f- L if e E ne rg y M at er ia ls A cq u. & P ro ce ss , A gr ic ul tu re , co ns tr uc t. & M an uf ., U se , E nd -o f- L if e, en er gy , T ra ns po rt C at eg or ie s of S ec to rs , F ac il it ie s, or U n it P ro ce ss es P ro du ct io n H ou se ho ld w as te m an ag em en t pr ac ti ce s M un ic ip al it ie s E ne rg y P ro du ct io n In du st ry , ut il it ie s, ag ri cu lt ur e, se rv ic es , e tc . Sp ec if ic S ec to rs , F ac il it ie s, U n it P ro ce ss or o r M at er ia ls C ap tu re d E le ct ri c ut il it ie s; f os si l fu el s (p ri m ar il y co al ), nu cl ea r, r en ew ab le r es ou rc es . In p u ts a n d O u tp u ts In cl u d ed C ap tu re s In pu t M at er ia ls or l an d us e, C ap tu re s O u tp u t P ro du ct s, C ap tu re s O u tp u t W as te s (r ec yc li ng ) P ro vi de s a D es cr ip ti on o f U ni t P ro ce ss es C ap tu re s In pu t M at er ia ls or l an d U se ( m at er i- al s in to l an df il ls a n d in ci ne ra ti on ), c ap tu re s O u tp u t W as te s (r ec yc li ng ) C ap tu re s O u tp u t P ro du ct s P ro vi de s a D es cr ip ti on of U ni t P ro ce ss es , C ap tu re s In pu t E ne rg y, C ap tu re s In pu t M at er ia ls o r la nd U se , C ap tu re s O u tp u t P ro du ct s, C ap tu re s O u tp u t W as te s o W n n r w > </> w C D w H > o od O O Se ct or N ot eb oo k s (U S E P A ) U ll m an n 's E nc yc lo pe di a O f In du st ri al C he m is tr y P ro ce ss P ro ce ss , M at er ia ls M at er ia ls A cq u . & P ro ce ss , A gr ic ul tu re , C on st ru ct . & M an uf ., E ne rg y M at er ia ls A q u c. & P ro ce ss , A gr ic ul tu re , C on st ru ct . & M an uf ., U se , E nd -o f- L if e, E ne rg y, T ra ns po rt M anuf ac tu ri ng an d A gr ic ul tu re C he m ic al P ro ce ss es 1 A gr ic ul tu ra l C he m ic al , P es ti ci de a nd F er ti liz er In du st ry ( 19 99 ), A gr ic ul tu ra l C ro p P ro du ct io n 1 In du st ry ( 19 99 ); A gr ic ul tu ra l L iv es to ck P ro du ct io n In du st ry ( 19 99 ); A er os pa ce In du st ry ( 19 98 ); (n ew ); A ir T ra ns po rt at io n In du st ry ( 19 97 ); D ry C le an in g In du st ry ( 19 95 ); E le ct ro ni cs a nd C om pu te r In du st ry ( 19 95 ); F os si l F ue l E le ct ri c P ow er G en er at io n In du st ry (1 99 7) ; i no rg an ic C he m ic al I nd us tr y (1 99 5) ; Ir on a nd S te el I nd us tr y (1 99 5) ; L um be r an d W oo d P ro du ct s In du st ry ( 19 95 ); M et al C as ti ng In du st ry ( 19 97 ); M et al F ab ri ca ti on I nd us tr y (1 99 5) ; M et al M in in g In du st ry ( 19 95 ); M ot or 1 V eh ic le A ss em b ly I nd us tr y (1 99 5) ; N on fe rr ou s 1 M et al s In du st ry ( 19 95 ); N on -F ue l, no n- M et al M in in g In du st ry ( 19 95 ); O il an d G as E xt ra ct io n 1 In du st ry ( 19 99 ); (n ew ); O rg an ic C he m ic al In du st ry ( 19 95 ); P et ro le um R ef in in g In du st ry (1 99 5) ; P ha rm ac eu ti ca l In du st ry ( 19 97 ); P la st ic R es in s an d M an -m ad e F ib er s In du st ry ( 19 97 ); P ri nt in g In du st ry ( 19 95 ); P u lp a nd p ap er In du st ry ( 19 95 ); R ub be r an d P la st ic I nd us tr y (1 99 5) ; S hi pb ui ld in g an d R ep ai r In du st ry (1 99 7) ; S to ne , C la y, G la ss a nd C on cr et e In du st ry (1 99 5) ; T ex ti le s In du st ry ( 19 97 ); T ra ns po rt at io n E qu ip m en t C le an in g in du st ry ( 19 95 ); W oo d | F ur ni tu re a nd F ix tu re s In du st ry ( 19 95 ) O ve r 80 0 ar ti cl es w ri tt en b y 30 00 e xp er ts , m or e th an 1 0, 00 0 ta bl es , 2 0, 00 0 fi gu re s, 16 m il li on w or d s - w h at ev er w ay y ou l oo k at it , U ll m an n 's E n cy cl op ed ia o f In du st ri al C h em is tr y of fe rs y ou a s tu p en d ou s am ou n t of in fo rm at io n in i nd us tr ia l ch em is tr y, p ro ce ss en gi n ee ri n g, m at er ia ls s ci en ce , en vi ro n m en ta l 1 ch em is tr y, f oo d sc ie n ce a n d b io te ch n ol og y P ro vi d es a D es cr ip ti on of U n it P ro ce ss es , P ro vi d es P ro ce ss F lo w D ia gr am s, C ap tu re s In pu t E ne rg y, C ap tu re s In pu t M at er ia ls o r L an d U se , C ap tu re s O u tp u t P ro d u ct s, C ap tu re s ou tp u t W as te s P ro vi de s a D es cr ip ti on of U ni t P ro ce ss es , P ro vi de s P ro ce ss F lo w D ia gr am s, C ap tu re s In pu t E ne rg y, C ap tu re s In pu t M at er ia ls o r L an d u se , C ap tu re s O ut pu t P ro du ct s, C ap tu re s O ut pu t W as te s | (C on tin ue d) C O o n o r4 h- l w n n r· w < w H O S I u> C O T ab le 5 .3 (c on t.) L ife c yc le in ve nt or y da ta s ou rc es , p re pa re d by J oy ce C oo pe r, U ni ve rs it y of W as hi ng to n. A va ila bl e at ht tp :/ /f ac ul ty .w as hi ng to n. ed u/ co op er js /D ef in it io ns /i nv en to ry sq ua re d. ht m ( ac ce ss ed J an ua ry 2 01 2) . D at ab as e T it le S of tw ar e an d T oo ls W it hi n C hi ef C le ar in gh ou se f or In ve nt or ie s A nd E m is si on F ac to rs (U S E PA ) T he G re en ho us e G as es , R eg ul at ed E m is si on s, a nd E ne rg y U si ng T ra ns po rt at io n (G R E E T ) M od el (A rg on ne N at io na l L ab or at or y) P ro ce ss es , F ac ili ty , Se ct or , o r M at er ia l- B as ed P ro ce ss a nd F ac il it y P ro ce ss A p p li ca b il it y M at er ia ls A cq u. & P ro ce ss , A gr ic ul tu re , C on st ru ct . & M an uf ., us e, E nd -o f- L if e, E ne rg y, T ra ns po rt E ne rg y, T ra ns po rt C at eg or ie s of S ec to rs , F ac il it ie s, or U n it P ro ce ss es In du st ri al , ag ri cu lt ur al an d tr an sp or ta ti on so ur ce s T ra ns po rt at io n an d fu el c yc le em is si on s Sp ec if ic S ec to rs , F ac il it ie s, U n it P ro ce ss or o r M at er ia ls C ap tu re d G as ol in e ve hi cl es ; F ed er al r ef or m ul at ed g as ol in e, C al if or ni a re fo rm ul at ed g as ol in e, E S O ; C ID I ve hi cl es : d ie se l) ; c om pr es se d na tu ra l ga s ve hi cl es ; b i- fu el , de di ca te d fu el ; D ed ic at ed L iq ue fi ed p et ro le um g as v eh ic le s, F le xi bl e- fu el ve hi cl es , E 85 , M 85 ; E le ct ri c ve hi cl es ; G ri d- co nn ec te d H E V s. C al if or ni a re fo rm ul at ed ga so li ne ; G ri d- in de pe nd en t H E V s; F ed er al re fo rm ul at ed g as ol in e, d ie se l S I ve hi cl es ; D ed ic at ed c om pr es se d na tu ra l ga s, d ed ic at ed li qu ef ie d na tu ra l ga s; d ed ic at ed l iq ue fi ed pe tr ol eu m g as , d ed ic at ed E 90 , d ed ic at ed M 90 ; C ID I ve hi cl es ; F ed er al r ef or m ul at ed g as ol in e, C al if or ni a re fo rm ul at ed d ie se l, d im et hy l et he r, F is ch er -T ro ps ch d ie se l; b io di es el , In p u ts a n d O u tp u ts In cl u d ed C ap tu re s O u tp u t P ro du ct s, C ap tu re s O u tp u t W as te s (A ir em is si on f ac to rs pr es en te d pe r un it o r ou tp ut ) C ap tu re s In pu t E ne rg y, C ap tu re s In pu t M at er ia ls o r la nd U se , C ap tu re s O u tp u t P ro du ct s, C ap tu re s O u tp u t W as te s I M O B IL E6 : V eh ic le a nd E ng in e E m is si on M od el in g So ftw ar e (U S EP A ) N at io na l E m is si on s In ve nt or y (N E I) A ir Po llu ta nt E m is si on s T re nd s D at a (U S 1 E PA ) Pr oc es s In di vi du al po in t so ur ce , fa ci lit y, co un ty le ve l T ra ns po rt E m is si on es tim at es f or ar ea , m ob ile an d ot he r so ur ce s T ra ns po rt at io n G ri d- in de pe nd en t H E V s. F ed er al r ef or m ul at ed ga so lin e, c om pr es se d na tu ra l g as , l iq ue fi ed na tu ra l g as , l iq ue fi ed p et ro le um g as , E 90 , M 90 , 1 re fo rm ul at ed d ie se l, di m et hy l e th er , fi sc he r- 1 T ro ps ch d ie se l, bi od ie se l; G ri d- co nn ec te d 1 H E V s. C al if or ni a re fo rm ul at ed g as ol in e, c om - pr es se d na tu ra l g as , l iq ue fi ed n at ur al g as , l iq - 1 ue fi ed p et ro le um g as , E 90 , M 90 , r ef or m ul at ed 1 di es el , d im et hy l e th er , F is ch er -T ro ps ch d ie se l,1 bi od ie se l; E le ct ric v eh ic le s, F ue l- ce ll ve hi cl es ; H yd ro ge n, m et ha no l. G as ol in e, e th an ol , c om - 1 pr es se d na tu ra l ga s In fo rm at io n fr om n um er ou s St at e an d lo ca l a ir 1 ag en ci es , t ri be s, a nd i nd us tr y on s ta tio na ry a nd 1 m ob ile s ou rc es th at e m it cr ite ri a ai r po ll ut an ts 1 an d th ei r pr ec ur so rs , a s w el l a s ha za rd ou s ai r | po ll ut an ts ( H A Ps ). C ap tu re s O ut pu t Pr od uc ts , C ap tu re s O ut pu t W as te s (A ir em is si on f ac to rs pr es en te d pe r un it of ou tp ut ) E st im at es o f a nn ua l e m is - si on s, b y so ur ce , o f a ir po llu ta nt s in e ac h ar ea of th e co un try , o n an an nu al b as is . C D o s o r > — I W n o f w < w Z H O I CO 134 LIFE CYCLE ASSESSMENT HANDBOOK Another challenge to using public databases is the difficulty in aggregating facility data for the precise group of facilities for which one may want data. Aggregating by SIC code generally groups facilities into too broad a category to be useful for LCA. An example is SIC code 3312, Steel Works, Blast Furnaces (including Coke Ovens), and Rolling Establishments. If one is interested in facilities that manufacture only steel tubing, SIC code 3312 is too broad of a category to provide useful information. The ability to aggregate on some smaller segment of industry would be extremely useful. Ideally, the names and addresses of facilities that produce the material or product of interest are known and the researcher can simply collect data for those facilities. This may be the case for LCA studies with a very narrow scope, but is generally not the case for LCA studies that include commodity products. 5.8 Creating Life Cycle Inventory using Economic Input/Output Data Economic Input/Output (EIO) analysis is an economic discipline that models the interdependencies of production and consumption between industries and households within a nation's economy. The input/output model divides an entire economy into distinct sectors and represents them in table, or matrix, form so that each sector is represented by one row and one column. The matrix represents sales from one sector to another. The economic input-output model is linear so the effects of purchasing $1,000 from one sector will be ten times greater than the effects of purchasing $100 from that sector. The data models the economic flows (in millions of dollars) of goods throughout the economy and includes a matrix of close to 500 industrial sec- tors. It shows how the output of one industry is an input to other industries. Today, almost all countries regularly compile IO tables to track their national accounts, although few are as detailed as the US model, which pro- vides data across approximately 500 sectors. With the growth of eCommerce, Table 5.4 Simplified economic input/output data available from the department of commerce. Economic Activity Agriculture Manufacturing Transportation Labor Etc.... Inputs to Agriculture 5 10 10 25 Inputs to Manufacturing 15 20 15 30 Inputs to Transport 2 10 5 5 Final Demand 68 40 0 0 Total Output 90 | 80 30 60 SOURCING LIFE CYCLE INVENTORY DATA 135 price information for most commodities is available through an on-line search. The US Department of Commerce's Bureau of Economic Analysis provides IO tables for the United States (http://www.bea.gov/industry/). Merging EIO with LCA referred to as EIOLCA, offers an alternative way to easily create LCI. To do so, the economic output for each sector is first cal- culated, and then the environmental outputs are calculated by multiplying the economic output at each stage by the environmental impact per dollar of output. The advantage of the economic input/output approach is that it quickly covers an entire economy, including all the material and energy inputs, thereby simplifying the inventory creation process. Its main disadvan- tage is the data are created at high aggregate levels for an entire industry, such as steel mills, rather than particular products, such as the type of steel used to make automobiles. Therefore, if the product being studies is representative of a sector, EIOLCA can provide a fast estimate of the complete supply chain implications. EIO-LCA methodology is a major research focus for the Green Design Institute at Carnegie Mellon University. Over the past 15 years, the group has investigated numerous products, services, and infrastructure systems using LCA as a fundamental component of analysis, leading field in EIO methodol- ogy and application, and produced an openly available on-line tool (http:// www.eiolca.net/methods.html). "Hybrid" models which combine the economic input/output model with process models have been proposed to utilize the advantages offered by both approaches (Sangwon and Huppes 2002; Hendrickson, Lave L. et ah 2006). See for example CEDA3.0 (http://www.iel.umn.edu/CEDA3_Users_Guide.pdf) 5.9 Global Guidance for Database Creation and Management A coordinated global effort to define and produce high-quality LCA data is required if LCA practice is to advance in the most resource-efficient manner. Further, a similar effort on data interchange is required to allow for the maxi- mum exchange of information among LCA practitioners. Only with wide- spread availability of LCA information will society be able to make efficient and effective decisions on policies and design options that will allow future generations to meet their own needs and aspirations. Guidance principles provide direction to users on selecting data that meet their needs, regardless of where an activity in a life cycle inventory (LCI) occurs. In addition, data developers and database managers need guidance on how to create datasets and operate databases, respectively, to provide exchangeable and fully documented datasets to users. Globally harmonized guidance will support an efficient allocation of resources, to ensure reliability and quality of data. Under the auspices of the Life Cycle Initiative, a joint effort between the United Nations Environment Programme (UNEP) and the Society of 136 LIFE CYCLE ASSESSMENT HANDBOOK Environmental Toxicology and Chemistry (SETAC), a process to lead to global agreement was established with the following vision: • To provide global guidance on the establishment and mainte- nance of LCA databases, as the basis for future improved dataset consistency and interlinkages of databases worldwide; • To facilitate additional data generation (including data specific to certain applications such as carbon and water footprint creation) and to enhance overall data accessibility; • To increase the credibility of existing LCA data through the pro- vision of guidance on the usability or fitness of data for various purposes; and • To support a sound scientific basis for product stewardship in busi- ness and industry, for life cycle-based policies in governments, and ultimately, to help advance the sustainability of products. In early 2011, a workshop was held in Japan for one week, bringing together 48 experts from six continents on both the provider and user sides to develop global guidance principles for LCA databases and the datasets they contain. This document, known informally as the Shonan Guidance Principles (UNEP/ SETAC 2011), are being widely promoted by UNEP and SETAC staff throughout the LCA community, targeting in particular database managers who are in a cen- tral role to support improvements in LCA datasets and database management. It is expected the guidance principles will be taken up worldwide and lead toward better
