Abstract
Chemical screening is often conducted using scoring and ranking methodologies in the United States.  Linked models accounting for chemical fate, exposure and toxicological effects are generally preferred in Europe and in product Life Cycle Assessment (LCA).  For the first time, a comparison is presented in this paper of two of the prominent but structurally different methodologies adopted to help screen and rank chemicals and chemical emissions data.  Results for 250 chemicals are presented, with a focus on twelve chemicals of interest in the United Nations Environment Programme’s (UNEP) Persistent Organic Pollutants (POPs) global treaty negotiations.  These results help illustrate the significance of described structural differences and to assess the correlation between the methodologies.  The scope of the comparison is restricted here to human health, although the insights will be equally useful in the context of the health of ecosystems.

Illustrating the current types of chemical screening and emissions comparison approaches, the relative significance of the scenario and structural differences of the Waste Minimization Prioritization Tool (WMPT) and the Toxic Equivalency Potential (TEP) methodologies are analyzed in this paper.  The WMPT facilitates comparison in terms of key physical-chemical properties.  Measures for Persistence, Bioaccumulation and Toxicity (PBT) are calculated.  Each PBT measure is scored and then these scores are added to provide a single measure of relative concern.  Toxic Equivalency Potentials (TEPs) account for chemical fate, multi-pathway exposure and toxicity using a model-based approach.  This model structure is sometimes considered to provide a less subjective representation of environmental mechanisms, hence an improved basis for screening.  Nevertheless, a strong relationship exists between the two approaches exits and both have their limitations.
 

Keywords: chemical, emission, screening, ranking, comparison, LCA, WMPT, TEP, POPs, PBTs
 

Introduction
Methodologies are in use by governments and industry [1, 2, 3, 4], or are being developed by international organizations such as the United Nations Environment Programme [5, 6], to help screen, rank and compare chemicals and emissions in product life-cycle assessment (LCA), process design and for regulatory chemical screening purposes.  While it is not the intention in such applications to yield explicit measures of risk associated with known impacts at any one given site, associated screening and comparison tools provide generic insights into the relative potential of a chemical, or an emission, to result in impacts of concern [7].  Unlike the more sophisticated, often chemical-specific studies conducted for chemicals such as DDT, dioxins, PCBs and mercury, tools are needed that provide a basis to screen or compare large numbers of chemicals.  For example, over 30000 new chemicals have been screened under Section 5 of the Toxic Substances Control Act (TSCA) [8].

In 1998, a workshop was conducted in Brussels to discuss issues related to methodological sophistication and comprehensiveness [9, 10].  With the aid of a hierarchical framework, see Table I, Pennington and Yue [4] similarly addressed the merits and limitations of available methodologies used for the comparison of chemicals and emissions in the context of regional-scale toxicological impacts.  Their hierarchy was based on the degree of representation and level of sophistication of associated environmental mechanisms (model structure).  It was, however, stated that the suitability of a given group of approaches in the hierarchy also depends on the relative environmental behavior of the chemicals considered, the quality of available data, the comprehensiveness of the model and the ability of more resource-intensive techniques to actually provide an improvement in discrimination.

As it is not possible to provide a detailed comparison of every approach in current use, we selected to study two well established and peer reviewed approaches.  The US EPA Waste Minimization Prioritization Tool (WMPT) [11, 12] and Toxic Equivalency Potentials (TEPs) [13, 14] have received considerable attention in the United States and can be considered to be classified into two of the higher hierarchical groups in Table I (scoring and ranking and model-based approaches, respectively).

The Waste Minimization Prioritization Tool (WMPT) relies heavily on a framework of expert judgment to identify chemicals, or emissions, of potential concern using key physical-chemical properties and associated pass/fail (cut-off) criteria.  WMPT was used as the foundation for developing US EPA’s Draft Resource Conservation and Recovery Act (RCRA) Persistent, Bioaccumulative and Toxic (PBT) List [5], a list of chemicals that, when final, will serve as the focus of voluntary US waste minimization program activities.  The WMPT was also proposed as a way to identify additional candidate chemicals for the US EPA’s PBT Initiative [15].  In terms of structure alone, the WMPT is a Group 3 comparison methodology.  This reflects a similar level in the hierarchy of many other chemical screening approaches in use in the United States, such as the assessment of new chemicals under Section 5 of TSCA [8] and Dupont’s chemical screening tool [16].

Toxic Equivalency Potentials (TEPs) are based on a generic version of the state-of-the-art model CalTox  [14, 17].  CalTox is one example of an integrated multimedia fate, multi-pathway exposure and toxicity model that was initially developed for use in regional human health risk assessments.  TEPs are a Group 4 comparison methodology in the hierarchy.  This higher classification reflects a perceived improvement in their representation of environmental mechanisms combined with a reduction in subjectivity in how the fate, exposure and toxicity parameters are combined [4].  Other Group 4 approaches include the comparison measures of Guinee et al. [18], based on the USES model [19] and recently revised by Huijbregts [20] for use in LCA, and the EUSES model for chemical screening within the European Union [21, 22].

Both the WMPT and the TEP types of screening methodologies account for chemical fate, exposure and toxicity, albeit using significantly different structures.  These tools are somewhat comparable in terms of objective and require the same minimum set of chemical input data.  Nevertheless the results will differ in terms of variability and three uncertainties: Scenario (associated with definition of the problem such as the region considered), Structural or Model (associated with the accuracy of relationships within the model in general and to what extent it represents true environmental mechanisms) and Parameter or Data (associated with the accuracy of both chemical and model parameters).

Parameter uncertainty and variability can be calculated using probabilistic techniques like Monte Carlo analysis [14, 23, 24].  Scenario and model uncertainties are not determined in such a parameter uncertainty analysis and are often ignored.  Only limited insights into these uncertainties are provided by evaluations using field data [23, 25, 26, 27, 28] or of components of the model [29].  Hence, although parameter uncertainty and qualitative differences associated with the structure provide some indications, the relative merits of different ranking approaches commonly remain unknown.  No "benchmark model" has been established for chemical screening and the comparison of emissions.

In the absence of suitable field measurement and techniques to evaluate available chemical screening methodologies, their comparison can provide useful insights.  In this paper, the relative significance of differences associated with the scenarios and structure of the WMPT and the TEP approaches are identified.  Results for 250 chemicals are presented to help illustrate the significance of described structural differences and to assess the correlation between the methodologies.1  Published results are first compared, illustrating the differences in data used as well as in the methodologies.  Twelve chemicals of interest in the United Nations Environment Programme’s (UNEP) Persistent Organic Pollutants (POPs) global treaty negotiations are highlighted and followed throughout the paper [5].  A detailed analysis is then presented of the fundamental differences between the approaches.

After the initial comparison of published results, the same input data are adopted for both approaches to eliminate chemical data related discrepancies.  Although the WMPT was subject to both internal review and public comment, data used to calculate the published TEPs were adopted due to their reputed level of peer review [13, 14].  The data for the twelve highlighted chemicals are presented in Table II.  It should be noted that this choice among the data sets does not necessarily guarantee higher quality and the actual level of review may prove to be lower in some cases.  For example, a higher proportion of measured fish BCF data were used in the published WMPT results for the set of 245 chemicals considered in this paper (40% in WMPT compared to 15% in CalTox).