I read the article by J. W. Owens entitled "Subjective Judgement" in Vol. 3, No. 1 edition of the J. of LCA. As usual, Dr. Owens writes an interesting article but I am left confused by the very restrictive statements related to human and ecological toxicological impacts. The statements related to subjectivity appear, at times, to be more demanding than in detailed risk assessments (also a very subjective decision making tool). Sometimes the normal level of subjectivity in risk assessment is not considered to be acceptable in LCA arena. Is there a justification for these higher demands that cannot often be achieved? For a tool that has no legal implications the level of subjectivity in LCA appears to receive more attention than may be necessary.

The author comments in the context of toxicological impacts that "... , beyond local aggregation is not plausible." However this aggregation is performed in risk assessments to determine the potential contribution of an emission to cause an effect in a specified region. Note, as in LCA, no actual effect is calculated (for a number of reasons summarized below). Furthermore, contrary to the authors statement, it appears that the loss of temporal information in the LCA inventory data limits the ability to perform risk assessment on a local scale to a greater extent than on a regional or global basis.

In the case of regional impacts, a release can be thought of in terms of its potential contribution to an effect. Hence, the contribution does not need to be a complete and single emission source. In the case of LCA, the contribution can be related to the fraction of a number of releases associated with a specified functional unit. Of course, the releases have to occur during similar time periods, for example one year. 

The occurrence of an actual effect in a region will depend on the temporal characteristics of the emissions, the extent of other releases in the region and the quantity of additional chemicals entering the region through advection. All chemicals released and entering by advection that have analogous toxicological effects have to be considered, unless there is evidence to suggest that the effects are not additive although similar. Hence, the prediction of actual impacts using LCA inventories, or any other emissions data, is typically not possible and only contributions can be estimated.

The potential contribution of a chemical to an effect in a region can be calculated using classical risk assessment paradigms. This was demonstrated using a deterministic approach in 1996 in the report by Guinee et al. "LCA Impact Assessment of Toxic Releases". *(Obviously the quality of data and hence the predictions can vary considerably.) The contribution potentials can be estimated and used to weight the inventory data. No actual impacts can be implied but the potential contribution to the regional cocktail is estimated on a relative basis using the minimal of scientific judgement currently possible.

Similarly, the statement of Dr. Owen that "the ratio of toxicity scores for different chemicals and systems is not likely to be proportional to their relative environmental loadings ..." requires clarification. It is not clear if the author is talking about the inability of LCA inventory data alone to reflect relative impacts or that the toxicity score of a particular chemical is not proportional to its potential contribution to an impact? If it is the latter statement then this is debatable, as the relative contribution of a specific chemical is approximately linearly proportional to the release magnitude, at least on a regional basis. Hence, contrary to the author�fs suggestion, regional potentials derived using common risk assessment approaches can be applicable for use in LCA.

One outstanding question is how to combine the resultant weighted inventory data for different chemical emissions. If the effect endpoints are similar then it is typical to add the results. However, to provide a "worst" case assessment for a given region, one may just add all the values. (Although this may not be the worst case if synergistic effects are considerable.) This summation is subjective and a value greater than one still does not indicate an actual impact. The value provides a relative indicator based on the sum of all the potential contributions associated with different effect endpoints. This is common practice in other impact/risk assessment arenas and no distinction is made between effects. A notable exception is between carcinogens and non-carcinogens, although there is suggestion that this practice may change.

Similarly, although more detail is gained by considering potential contributions on a regional scale, the decision maker will probablysum the values for the different regions to facilitate a comparison of the different life-cycles. Again this may not indicate actual effects but it will provide an indication of the overall contributions of one life-cycle to facilitate comparison with another in the category of toxicological impacts. Furthermore, the lack of distinction between areas will avoid discrimination against processes in less pristine regions of the world. This has both good and bad connotations.

The alternative to simple aggregation may be to consider background concentrations of all the chemicals in each region with a similar effect endpoint as well as the temporal data for each emission, etc. Unfortunately, this data is not typically available, even for most "detailed" risk assessments. So, the decision will typically have to be made on the basis of aggregated data across different effect endpoints and regions, unless the application of LCA is to be restricted. As this is the current "state-of-the-art" in many risk/impact assessment domains, not just in LCA, it is hard to see why this practice should be limited on this basis. However, there is obviously room for improvement in the LCA approach and the methodologies used to develop potentials. These advancements may include the inclusion of temporal information to enable the consideration of contributions occurring over different time spans.

* Guinee J., Heijungs R., van Oers L., van der Meent D., Vermeire T., Rikken M., LCA Impact Assessment of Toxic Releases, Ministerie van Volkshuisvesting, Ruimtelijke Ordening en Milieubeheer, May 1996
 

Dr. David W. Pennington