The elusive consensus on life-cycle assessment

Scientific consensus is urgently needed on the "impact assessment" stage of life-cycle assessment (LCA) - where emissions and other impacts are aggregated into a single score - according to LCA practitioners at a Society of Environmental Toxicology and Chemistry (SETAC) conference in April. About ten methods are currently in use or being developed and the findings of an LCA can be turned on its head depending on which one is used - leaving the technique still open to abuse.

By and large, LCA research has moved on from the "life-cycle inventory" (LCI) stage, where data on resource consumption, pollutant releases and waste generation associated with a product or process are obtained and tabulated. Complex issues, such as allocation techniques for multi-product processes and the setting of system boundaries, have generally been resolved.

Code of practice
The next stage is the "impact assessment" component. This is "perhaps the most difficult and least developed", according to an LCA Sourcebook from the Society for the Promotion of LCA Development (ENDS Report 222, pp 27-28 ).

Impact assessment has four steps, according to SETAC's recent code of practice (ENDS Report 224, p 30 ), although some methods combine or skip some steps. Data on, say, emissions from the LCI are first "classified" into a number of impact categories and then "characterised" - the aggregation of emissions within each impact category. The scores in each category are then "normalised" to iron out differences in orders of magnitude and units. Finally, the impact categories are aggregated into one final score. This "valuation" stage is "the trickiest step in a tricky process", says the LCA Sourcebook.

The first classification step is largely defined. There are two common approaches. In the first, emissions are aggregated by the medium into which they are released, while in the second they are aggregated according to their potential contribution to specific environmental effects.

SETAC has opted for the latter approach. Ten impact categories have been identified: depletion of abiotic resources, depletion of biotic resources, global warming, ozone depletion, human toxicity, ecotoxicity, photochemical oxidant formation, acidification, eutrophication and land use. The list is not definitive, SETAC acknowledges.

The characterisation and valuation stages involve aggregation. This was the main theme of SETAC-Europe's fourth conference in Brussels during April.

Equivalency factors
At the characterisation level, much attention is presently being paid to the development and use of "equivalency factors" for the different impact categories, following a methodology developed by the Centre of Environmental Science at Leiden University (CML) in 1992.

For some categories, this is fairly straightforward. Emissions of carbon dioxide, methane and other greenhouse gases, for example, can be combined using their "global warming potentials". Similarly, ozone depleting substances can be combined using ozone depletion potential.

But for other categories, such as human and ecotoxicity and biodiversity, the task is far more complex. There are at least nine current approaches for ecotoxicity alone - "too many divergent paths," CML's Jeroen Guine told the meeting.

CML has opted for a non-site specific approach where "potential" rather than "actual" impacts are quantified. But some LCA practitioners, including Jos Potting of Utrecht University in the Netherlands, argue that account should be taken of the specific pollutant receptor. For example, releases of SO2 and NOx have a far greater impact in regions that suffer acidification problems. CML's approach should be modified or extended for this reason, she said, "because it will select the wrong improvements or wrong products."

Valuation techniques
But it is at the valuation stage where there is the widest divergence of ideas. Comparing global warming with ozone depletion and other impacts is like comparing apples and oranges and cannot be done, argue the sceptics. SETAC has formed a working group on the issue to work towards a common approach.

There are some ten valuation methodologies. The most basic and least scientifically justified of these - and the one used most commonly by novices to LCA - is simply to add the emissions together without any regard for their polluting effect. Some of the more scientific methods are outlined below.

  • Swiss critical volumes: This was one of the first semi-aggregation techniques, and was developed by the Swiss Environment Ministry, BUWAL. The critical volume of each emission is the volume needed to dilute it to the concentration permitted by a regulatory standard. The values for each air pollutant are then added to give a total critical volume for air emissions, and similarly for discharges to water.

    However, this approach suffers from a lack of transparency because it combines the three steps of impact assessment, and therefore falls outside the SETAC framework. It is also not applicable to substances for which no regulatory standards have been set, is not strictly scientifically based, and refers only to the situation in Switzerland and Germany.

  • Swiss Eco-points: This "ecoscarcity" approach was also developed by BUWAL and, unlike the earlier version, gives a final single score. Eco-points are calculated according to societal goals - namely government objectives.

    The weighting given to each emission is the actual emission of a pollutant within a given area as a proportion of the maximum permissible load based on policy targets. The higher the actual level compared with the policy level, the higher the weighting given to that emission.

    The method suffers many of the same drawbacks as the original BUWAL approach in that it combines classification with valuation, and therefore some transparency is lost. However, it could be adapted to SETAC's framework if policy targets are set for each of SETAC's impact categories rather than each emission.

  • Dutch Environmental Performance Indicators (EPIs): This scheme is currently being developed by CML, as an extension of its classification\characterisation technique. It is similar to the Swiss "eco-point" method but bases weightings on sustainable, rather than policy, levels for each impact category. CML is attempting to calculate what the "sustainable" levels are for each category, and this is not easy.

    CML's approach is again global, without any attempt to incorporate site-specific considerations. But a project funded by the Nordic Council of Ministers is attempting to incorporate some degree of actual impact. For some categories, weighting factors will be evaluated under different site scenarios and presented in the final results as a range.

  • Difference between target and present levels: For both the "policy" and "sustainable" level systems, a weighting can give even greater bias towards impact categories where the actual level is far from the target level if it is based on the difference between the present and target levels as a proportion of the target level. However, when the actual level is less than the target level, then no weight is given, and this may not be environmentally justified.

  • Environmental Priority Strategies (EPS): The other main line of thinking is based on monetary parameters. The EPS approach has been developed by the Swedish Environmental Research Institute (IVL). In this the weightings used are based on the costs which society is willing to pay to restore the environment.

    Although this approach may reflect the actual preferences of society, its most fundamental drawback is that the cost of, say, preventing an emission may bear no relation to the damage it is causing. To some extent it also combines the various stages of LCA, and therefore lacks clarity.

    Alternatively, the weighting can be based on the investment society has to pay to meet policy objectives. This has been developed by the Tellus Institute in the USA. A drawback of this approach is that when actual levels are below policy levels, the effect is assumed to be absent. The weighting used also depends on the characteristics of the available technology and processes to be optimised. Like the EPS approach, it also has no direct relationship with effects.

  • Panel judgement: A common approach is to use a panel of "experts" to make judgements on the relative weightings to be given to each impact category. This can reflect the preferences of society, but most experiments have proved that the results are non-repetitive.

    Decision theory techniques can be employed to make value judgements more explicit, but only a few LCAs have been conducted along these lines. One trial was Landbank's analysis of zeolite and phosphate detergents, which used the Delphi technique (ENDS Report 228, pp 29-31 ).

    In the EC's eco-labelling programme, these value judgements are made through negotiations between officials, industry, and environmental and consumer groups from many or all of the Member States. Accordingly the end-results may well be based on politics rather than science. The European Commission is developing a common LCA methodology for eco-labelling to ensure that LCAs are carried out using similar techniques. But this may not go as far as the valuation stage.

    Comparing weightings
    Some of the weightings derived using these methods have been compared by the Dutch TNO Study Centre for Environmental Research in Delft. The weightings, relative to the "dispersion of toxic substances" impact category, are shown below for Dutch conditions for four categories. TNO warns that these are tentative results, but concludes that "weighting could have a similar uncertainty as the impact values being weighted."

    Each method gave very different weightings - sometimes by more than one order of magnitude. For example, the weighting for climate change varied by a factor of 65, and for eutrophication by a factor of 13.

    The differences between the climate change weightings are readily explained. The lowest weighting is under the system based on actual investment, indicating that investment in CO2 abatement is lagging behind policy goals. Conversely, climate change gains importance in the "sustainable level" approach.

    Conflicting results
    Most LCAs use only one type of impact assessment method. But a consultancy, stfold Research Foundation, has compared the production and distribution - but not combustion - of diesel, petrol, and petrol containing MTBE for Statoil using three methods: the Swiss Eco-points assessment method based on both Dutch and Norwegian targets, and the EPS method.

    Overall, the analysis suggested that petrol containing MTBE has the highest environmental impact, while diesel has the lowest. However, under the EPS system there was little to choose between normal petrol and petrol containing MTBE, while diesel showed far less of an advantage than under the Eco-points method.

    Looking at the production of petrol containing MTBE in detail, VOC emissions were given the highest weighting according to the Eco-points method based on Norwegian targets. The Dutch Eco-points system gave a higher weight to SO2 but a lower weight to CO2 due to differences in priorities between the Netherlands and Norway. Under the EPS "willingness to pay" approach, consumption of fossil fuel moves up the priority scale - from fourth under Eco-points to second under EPS.

    The weightings send very different messages to anyone interested in, say, planning environmental improvements. Under the Dutch Eco-points system, distribution impacts contribute most to the total score, whereas in the EPS approach it is crude oil refining. The high score in the latter is due to fossil fuel consumption at the refinery. The EPS system, warn the consultants, "values fossil fuel consumption so high that all other impact categories lose weight and importance."

    The packaging consultancy PIRA, which has developed an LCA software package, leaves it up to commissioning organisations to use their own weighting factors. For example, the Alliance of Beverage Cartons and the Environment weighted inventory data according to some very subjective parameters that rewarded low fossil fuel consumption (ENDS Report 230, pp 25-26 ). The result was that cartons came out looking better in environmental terms than plastic bottles.

    PIRA's justification for not incorporating weightings at this stage is that "it is unlikely that an entirely objective standard valuation method will be is a subjective choice between dissimilar parameters." However, "a lot of progress has been made," remarked CML's Jeroen Guine at SETAC's meeting. "We now know what we are talking about...but there is no doubt about it, a lot of work is still necessary."

    The call for consensus was loud in Brussels. If this is forthcoming, it could be a forerunner for more scientifically-based aggregation techniques in other areas, such as eco-labelling, environmental reporting, and in assessing best available techniques under, say, integrated pollution control.

    This would give some users what they want - a single environmental score for a process or a product. But single scores can mask the underlying environmental problems in each area. More fundamentally, impact assessment "suggests a scientific precision which cannot hold true," warned Mr Guine. Qualitative aspects are difficult - some would say impossible - to handle.

  • Please sign in or register to continue.

    Sign in to continue reading

    Having trouble signing in?

    Contact Customer Support at
    or call 020 8267 8120

    Subscribe for full access

    or Register for limited access

    Already subscribe but don't have a password?
    Activate your web account here