Acknowledging USEtox Documentation Clarity

Before examining a critical source of misunderstanding in how USEtox results are interpreted and communicated, it is important to explicitly acknowledge that the USEtox documentation is technically clear and honest about the comparative nature of the model. The official documentation consistently uses qualifying language—"comparative," "relative," "potential"—and explicitly states that USEtox provides "substance-specific measures of relative impact potential" for "comparative assessment contexts." The USEtox team has been scientifically rigorous in framing their model's purpose and limitations.

However, despite this clarity in the technical documentation, there exists a systematic source of misinterpretation that arises from the terminology itself: specifically, the use of "PAF" (Potentially Affected Fraction) and "PDF" (Potentially Disappeared Fraction) in the units of USEtox characterization factors. These terms carry established meanings in ecotoxicology and risk assessment that directly suggest quantification of real environmental impacts—meanings that do not align with what USEtox actually calculates.

Figure 1 shows the official USEtox 2.0 framework diagram that appears in the model documentation and is routinely used to explain the model structure to practitioners, researchers, and policymakers.

This widely-circulated framework diagram is particularly revealing because it shows how the model's outputs is presented:

For Ecosystem Toxicity (left side, green):

·      Midpoint indicator: "Potentially affected fraction of species"

·      Endpoint indicator: "Potentially disappeared fraction of species (damage on ecosystem quality)"

A practitioner viewing this diagram would naturally conclude: "USEtox measures the fraction of species affected and the fraction that will disappear, just as it measures disease incidences and disability-adjusted life years for humans (this unit will be discussed in a future article)."

For the vast majority of users who encounter this diagram but never read the full 208-page technical documentation, this visual representation becomes their primary understanding of what USEtox calculates.

The terminology—"potentially affected fraction," "potentially disappeared fraction," "damage on ecosystem quality" creates an almost irresistible interpretation: USEtox quantifies real ecological impacts from chemical emissions.

This terminological issue is not merely semantic. The language used in impact assessment frameworks shapes how results are understood by practitioners, communicated to decision-makers, and presented to the public. When a metric is expressed in units of "PAF·m³·days/kg," it strongly suggests—regardless of disclaimers in documentation that many users never read—that the score represents the fraction of species that will actually be affected by a product's chemical emissions. This perception, though technically incorrect, is a natural interpretation of the terminology chosen.

For ecotoxicologists trained in risk assessment, encountering "PAF" used to describe a comparative index rather than an actual affected fraction often triggers skepticism or rejection of the LCA approach entirely—a professional friction that undermines the collaborative potential of Safe & Sustainable by Design frameworks.

Furthermore, this terminological confusion is particularly consequential when USEtox is promoted and applied beyond its original product-comparison context, specifically in global chemical footprinting and planetary boundaries assessment for chemical pollution. In these applications, the suggestion that scores represent real ecosystem impacts (rather than comparative persistence-toxicity metrics) fundamentally mischaracterizes what is being measured.

This article examines how the terminology creates systematic confusion, the gap between what "PAF" means in traditional ecotoxicology versus what it represents in USEtox, and the implications when this confusion propagates through LCA practice and into policy applications.

What "PAF" Means in Ecotoxicology

In traditional ecotoxicology and environmental risk assessment, PAF (Potentially Affected Fraction) has a specific, well-established meaning:

PAF: Fraction of species in an ecosystem for which exposure concentration 'C' exceeds their effect threshold

Where: C = actual environmental concentration [µg/L or mg/kg]

This is derived from Species Sensitivity Distributions (SSDs), where toxicity data for multiple species are fitted to a statistical distribution. The PAF at concentration C represents the cumulative fraction of the species distribution for which that concentration would cause effects. The PAF can be at HC5, HC10 or HC20 like in USEtox.

Key characteristics of ecotoxicological PAF:

• Concentration-dependent: PAF is a function of actual (or predicted) environmental concentration

• Threshold-based: Usually HC5 in RA derived from effect concentrations (EC10, EC20, NOEC, etc.) for multiple species. USEtox uses HC20. 

• Interpretable as real impact: If PEC = 10 µg/L and HC5 = 10 µg/L, this means approximately 5% of species in the exposed ecosystem would experience effects at that concentration

• Used for setting quality criteria: Environmental quality standards (e.g., Water Framework Directive) use PAF = 0.05 (5% affected) as an acceptable protection level

What "PAF" Means in USEtox Units

In contrast, the "PAF" appearing in USEtox characterization factors is a mathematical construct that does not represent the fraction of species affected in any real ecosystem:

USEtox characterization factor:

CF [CTUe/kg] = FF [days] × XF [dimensionless] × EF [PAF·m³/kg]

Where EF is derived as:

EF = 0.2 / HC20   [PAF·m³/kg]

HC20 = Hazardous Concentration affecting 20% of species [kg/m³]

What this "PAF" actually represents:

The Effect Factor (EF) containing "PAF" in its units is calculated from the HC20, which itself is derived from Species Sensitivity Distributions.

• It is used as a toxicity potency parameter: The inverse of HC20, scaled by 0.2

• Integrated over time and space: The final "PAF·m³·days" represents time-integrated, volume-integrated persistence-toxicity potential

• No actual concentration calculated: The metric does not calculate what concentration will occur or whether any effect threshold will be exceeded

What this unit actually means:

This is not "the fraction of species affected for X days in Y volume of water." It is a comparative index that combines:

• Persistence (how long chemical mass remains in environment)

• Bioavailability correction (what fraction is truly dissolved)

• Toxicity potency

The 'PAF' here is simply 0.2/HC20 - a mathematical scaling factor, not a measured or predicted ecological response.

The "PAF" in the USEtox units is borrowed ecotoxicological terminology that does not have the same interpretation as in concentration-response assessment.

The Semantic Gap: PAF as "Impact" vs. PAF as "Potency Index"

The fundamental issue is that the same three-letter acronym is used for two completely different concepts:

Ecotoxicological PAF (Real Impact)

·      Input: Actual or predicted environmental concentration

·      Output: Fraction of species expected to experience effects at that concentration

·      Question answered: "What proportion of species will be harmed?"

·      Interpretation: Direct measure of ecological impact at a given exposure level

·      Use case: Setting environmental quality standards, evaluating whether PEC is acceptable

USEtox "PAF" (Comparative Potency)

·      Input: Mass emitted (inventory) to unknown volume of water

·      Output: Time-volume-integrated persistence-toxicity index

·      Question answered: "Which chemical has higher persistence × toxicity potential?"

·      Interpretation: Comparative ranking metric, not a prediction of actual ecological effects

·      Use case: Comparing lifecycle impacts between product alternatives

The critical difference: One is a dose-response relationship (ecotoxicology PAF), the other is a persistence-toxicity product (USEtox "PAF").

The Misleading Parallel

When a LCA practitioner sees:

·      Product A ecotoxicity: 150 PAF·m³·days

·      Product B ecotoxicity: 75 PAF·m³·days

The terminology naturally suggests: 

❌ "Product A affects species across twice the volume-time compared to Product B" 

❌ "Product A causes ecosystem impacts lasting twice as long" 

❌ "The PAF from Product A is twice as large"

What it actually means:

✓ "Product A has twice the persistence-toxicity potential index" 

✓ "In a comparative LCA, Product A ranks worse on this metric" 

✓ "No information about whether either product causes actual ecosystem effects"

The number "150" does not mean 150 m³ of water with species affected. It does not mean 150 days of effects. It is a dimensionless comparative index that happens to have units that look like they describe real spatial-temporal impacts.

The "Endpoint/Damage" Terminology Compounds the Problem

The confusion is further compounded by LCA's distinction between "midpoint" and "endpoint" (or "damage") indicators:

Midpoint Level: "Ecotoxicity Potential"

Unit: CTUe [PAF·m³·days/kg emitted]

·      Described as: "Potentially affected fraction of species integrated over volume and time" (this is what is said in Figure 1)

·      Conceptually positioned as: An intermediate indicator

Endpoint Level: "Ecosystem Quality Damage"

Unit: CDUe [PDF·m³·days/kg emitted]

Where: PDF = Potentially Disappeared Fraction = 0.5 × PAF

·      Described as: "Potentially disappeared fraction of species integrated over volume and time"

·      Conceptually positioned as: Damage to the area of protection "Ecosystem Quality"

The problematic implication of "damage" terminology:

In LCA methodology, "damage" or "endpoint" indicators are meant to represent actual harm to areas of protection (human health, ecosystem quality, resource availability). 

For ecotoxicity, calling PDF·m³·days "ecosystem quality damage" strongly suggests:

·      This is a measure of actual harm to ecosystems

·      Species are actually disappearing (the language of "disappeared fraction")

·      Ecosystem quality is actually being damaged

Yet, as established earlier, USEtox does not calculate:

·      Whether any concentration threshold is exceeded

·      Whether any observable ecosystem effect would occur

·      Whether any species would actually be affected, let alone disappear

The mathematical transformation (PAF → PDF = 0.5 × PAF) does not change the fundamental nature of the metric. It remains a persistence-toxicity index, not a measure of real ecosystem damage. 

Recommendations: Toward Clearer Terminology

To reduce the systematic misinterpretation while preserving USEtox's valid comparative purpose, a terminological reform could help:

Neutral Comparative Index Units

Replace:

·      CTUe [PAF·m³·days/kg]

·      CDUe [PDF·m³·days/kg]

With:

·      Ecotoxicity Potential Index [EPIₑ]

·      Ecotoxicity Damage Index [EDIₑ]

Units: Dimensionless or arbitrary units (e.g., "index points" or "toxicity-persistence units")

Advantage: Immediately clear that this is a comparative ranking metric, not a quantification of real impacts

Disadvantage: Breaks continuity with existing database, requires recalibration of practitioner understanding

Conclusion: Honest Tools, Misleading Language

To be absolutely clear: USEtox is a scientifically sound comparative assessment tool, and its developers have been technically honest in their documentation. The model does what it claims to do when properly understood within its validated domain.

However, the terminology chosen to express USEtox results systematically creates misunderstanding by borrowing language (PAF, PDF, damage, affected fraction, disappeared fraction) from risk assessment and ecotoxicology where these terms describe real, observable impacts rather than comparative persistence-toxicity indices.

This is not a minor semantic quibble. It has real consequences:

1.        LCA practitioners misinterpret absolute values, thinking they quantify real ecosystem impacts

2.        Substitution decisions are made based on misunderstood changes in comparative indices rather than actual environmental risk reduction

3.        Global chemical footprint assessments are presented as quantifying real planetary-scale ecosystem damage when they measure aggregate persistence-toxicity indices

4.        Policymakers and the public are led to believe we are measuring actual chemical impacts when we are calculating comparative rankings

5.        Official framework diagrams present ecosystem metrics in parallel with genuine health impact metrics, reinforcing the misperception that they are analogous types of measurements

The fundamental problem: You cannot use the language of impact assessment (PAF, PDF, damage, affected fraction, disappeared fraction) while delivering a comparative ranking tool, and expect systematic correct interpretation.

This terminological issue must be addressed—either by reforming the nomenclature to accurately reflect what is being calculated, or through sustained educational efforts to ensure every user understands that "PAF" in USEtox means something fundamentally different from "PAF" in ecotoxicology. Until then, even the most technically sound comparative tool will be systematically misused as if it quantifies real environmental impacts.