Normalization: new developments in normalization sets
This is the second in a series of articles where PRé's own consultants explain a step in the LCIA process and the way their own expertise fits into that step. Tommie Ponsioen is a technical consultant at PRé, with expertise in LCA methodology and database development. This article explains normalisation, the third step in LCIA, and describes new developments in the field.
Why comparing your product’s impacts to Average Joe’s makes sense
The environmental impact scores of life cycle assessments are often presented in units that are difficult to grasp, such as kg CO2 equivalents or CTUh. One way to make interpreting such scores easier is to normalize them: dividing your scores by a reference situation’s scores. This reference situation could be one person’s – Average Joe’s – share of all emission and resource use in the world during one year. Normalization converts complicated units into fractions of Joe’s scores per impact category.
The scores for climate change, human toxicity, and many other impact categories are all compared to the annual impact Average Joe has and expressed in fractions .
As an example to illustrate how this works, I’ll show the normalization of the carbon footprint of milk. One liter of milk has an impact of about 1 kg CO2eq. For comparison, global greenhouse gas emissions in 2010 were about 7 ton CO2 equivalents per person. Therefore, one liter of milk has an impact of 1/7000 = 0.00014 person-year per liter. Average Joe drinks about 50 liters of milk per year. Hence, the milk consumption of Average Joe is equivalent to an impact of 0.007 person-year, or 0.7% of climate change impact of an average person in a year.
This exercise can be done for any impact category at midpoint level (e.g. equivalence factors) or endpoint level (e.g. damage to human health), and for any average person at country, region or world level.
Product comparison: another good reason for doing normalization
The example shows that normalization gives you an idea of an impact’s magnitude. Another reason for doing normalization is to compare the impact of two products with a multi-impact assessment method. With normalization you can add up the results of the different categories because they have the same unit. It is recommended to give each category a different weight, even though weighting is, by definition, arbitrary. The next article in this series will elaborate on this.
Normalization made easy in LCA software
A few product impact assessment methods include one or more normalization sets to choose from in LCA software, each covering different regions and years. The table below provides an overview.
As you can see, not all methods contain the same normalization sets and some sets are rather outdated. Even so, it is not possible to apply normalization sets from one method to another. One of the problems with that is that normalization sets are calculated from normalization inventories – a record of all emissions and resource extractions in a region in a year– that are often specific to a method. Since the coverage of emission and resource use differs per method, normalization inventories, and the resulting normalization sets, cannot be applied to other methods.
More complete normalization inventories are being developed
There are several groups of scientists and consultants working on developing more complete and accurate normalization inventories. This is possible partly because statistics on emissions and resource use from international organizations, such as the United Nations and the European Commission, are becoming more complete. The new developments are also stimulated by the increasing demand for normalization inventories as new multi-impact methods are being developed and existing methods are combined.
There are three developments I want to mention in particular.
- A new normalization set was published recently for the ILCD methods compilation with a reference of the EU27 in the year 2010. It will be available in LCA software soon.
- The IMPACT World+ method with accompanying normalization is about to be released. The method has characterization factors for a large range of environmental impact categories at the midpoint and endpoint levels.
- The LC-IMPACT method is still under construction. The method is mainly based on endpoint modeling and also covers a large range of environmental impact categories. It will include normalization references at the European and global levels.
Having a complete normalization inventory is crucial
All three methods mentioned above have a large scope with many environmental impact categories, which require a large coverage of emitted and extracted substances. This means that the normalization inventories need to be extensive as well. Because the normalized impact of a product is equal to the product impact divided by the reference situation impact, lack of data can severely affect the normalized product impact assessment in both directions. If a substance with a significant contribution to the impact of the reference situation is not included in the impact of the analyzed product, the assessed product impact would be estimated too low. And if the reference situation underestimates the impact of a specific category, such as carcinogenic human toxicity, the normalized assessed product impact would overestimate toxic emissions in the life cycle of a product. As a consequence, the contribution of this impact category to the overall environmental impact becomes incorrectly dominant.
Filling the gaps can be a challenge
The available statistics on emissions and resource extractions are extensive, but filling in the essential remaining gaps with all sorts of estimation techniques is still a lot of work. For example, the list of toxic substance emissions to air, water and soil is very long, which makes it impossible for governments to monitor and report them all in detail. These figures need to be estimated with scientific models, which – no matter how sophisticated – nevertheless include some important assumptions. The extraction of metals is another area where statistics are incomplete because not all companies want to make their production figures public. Filling in the gaps is even more challenging in this case because there are virtually no scientific estimation models available.
For some impact categories, on the other hand, it is much easier to fill in those gaps. Some examples:
- Carbon dioxide has the highest contribution to climate change and is monitored in detail, and the other two important greenhouse gases, nitrous oxide and methane, are estimated using equations and parameters supplied by the UN. The United Nations Framework Convention on Climate Change (UNFCCC) publishes the results per country on their website each year.
- Fossil resource extractions are also well monitored and documented, for example by the UN (UN data), the European Commission (Eurostat), U.S. Energy Information Administration (EIA), and the International Energy Agency (IEA).
- Particulate matter emissions – primary emissions and precursors – are reported by the European Environment Agency (EEA) and by the United States Environmental Protection Agency (US EPA), but estimates are still needed for other countries.
- Detailed land use data for land occupation and transformation impact categories can be found in the CORINE land cover database of the EEA for Europe and the World Resource database of the World Resources Institute (WRI), the World Bank and the UN for the rest of the world.
Understanding the sources of regional environmental impacts
Data on reference situations are not just useful for normalization, but also for analyzing environmental impacts in different countries or regions over the years. This is especially interesting when using an endpoint method, which enables the identification of important emission sources or resource uses related to human health, ecosystems or resources. Such analyses also help us understand the value of normalized impact scores. Therefore, it is good news for everyone interested in normalization or comparative impact analysis that several new normalization inventories and analyses will be published in scientific and policy journals in the coming years.
Read about other LCA steps, each explained by an expert:
Tommie Ponsioen
Tommie worked for PRé as a Technical Consultant from 2012 until 2015. As a part of the Consultancy Team, he worked with databases and methods. Tommie collaborated in projects such as Prosuite and improving the ReCiPe method.