This mutual skepticism between the LCA and risk assessment (RA) communities is understandable—each discipline has developed its own language, its own priorities, and its own tools. But here's something that might surprise both communities: USEtox (the consensus model for ecotoxicity characterization in LCA) and EUSES (the European System for chemical risk assessment) are built on the same scientific foundation.

More specifically, both rely on SimpleBox, the multimedia environmental fate model developed in the Netherlands. Both use the same physicochemical properties. Both model the same environmental compartments. And both could—theoretically—draw from the same ecotoxicity databases generated under REACH.

Yet, they produce fundamentally different outputs: USEtox delivers relative impact potentials (Characterization Factors in CTUe), while EUSES calculates absolute concentrations (PECs in µg/L) for comparison with safety thresholds.

In the context of Safe and Sustainable by Design (SSbD), this isn't a problem—it's an opportunity. The European Commission's SSbD framework explicitly calls for integrating both safety (regulatory risk assessment) and sustainability (life cycle thinking) pillars. Understanding how these two tools complement each other, despite their different purposes, is crucial for maximizing the value of expensive REACH toxicological data.

This article aims to bridge the gap between LCA practitioners and risk assessors by explaining:

• How USEtox and EUSES use the same underlying complexity (SimpleBox)
• Why they produce different outcomes despite this shared foundation
• How both approaches can work together in the SSbD framework

Let's start by examining what these tools have in common.

The Shared Foundation: SimpleBox and Environmental Fate Modeling

What is SimpleBox?

SimpleBox is a multimedia environmental fate model that simulates how chemicals distribute and degrade in the environment. Originally developed at the Dutch National Institute for Public Health and the Environment (RIVM) in the early 1990s specifically for risk assessment purposes, SimpleBox treats the environment as a series of interconnected "boxes" (compartments) where chemicals can:

• Partition between air, water, soil, and sediment phases
• Degrade via photolysis, hydrolysis, biodegradation
• Transport between compartments through advection, diffusion, and deposition
• Bioaccumulate in organisms

Both EUSES (which implements SimpleBox 3.0) and USEtox (which evolved from SimpleBox principles) rely on this fundamental modeling architecture. Importantly, SimpleBox was first implemented for regulatory risk assessment in EUSES, and the LCA community later adopted and adapted these principles for USEtox development in the mid-2000s. This historical sequence matters: USEtox didn't invent multimedia fate modeling—it adapted proven risk assessment methodology for comparative life cycle purposes.

Common Requirements: The Same Input Data

To run either model, you need the same set of physicochemical properties:

These properties aren't arbitrary choices—they're the fundamental parameters that govern how any organic chemical behaves in the environment, regardless of whether you're doing risk assessment or life cycle impact assessment.

Common Compartments and Processes

Both models simulate the same environmental system:

Compartments:

• Atmospheric air (with aerosols)
• Freshwater bodies
• Marine water
• Agricultural soil
• Natural soil
• Freshwater and marine sediments
• Vegetation (with some differences in implementation)

Physical-chemical processes:

• Advective transport (wind, water flow)
• Diffusive exchange between compartments
• Wet and dry deposition from air
• Sedimentation and resuspension
• Soil-air exchange
• Plant uptake
• First-order degradation in all media

The Same Modeling Complexity

Here's a critical point that both communities should appreciate: Neither USEtox nor EUSES is "simpler" than the other when it comes to environmental fate modeling.

Both implement multi-compartment mass balance equations. Both require understanding of:

• Fugacity-based equilibrium partitioning
• Environmental transport coefficients
• Degradation kinetics
• Spatial nesting of boxes

The complexity is equivalent. The core difference isn't in the sophistication of the fate modeling—it's in how the models are used and what questions they're designed to answer.

The Divergence: Different Scales, Different Emissions, Different Outputs

Now let's examine where USEtox and EUSES diverge, despite their common foundation.

Level 1: Exposure/Fate - Different Spatial Scales and Emission Scenarios

EUSES/Risk Assessment: Real-World Scenarios

When a risk assessor uses EUSES, they're modeling a specific, real-world situation:

User inputs:

• European tonnage: 5,000 tonnes/year

• Local site tonnage: 150 tonnes/year

• Emission pattern: 300 days/year, continuous

• Fraction to wastewater: 0.1%

• Industrial category and use pattern

EUSES then calculates:

1. PECcontinental: Background concentration across Europe based on total tonnage distributed over the entire continental freshwater volume (~3.24×10¹⁴ m³)
2. PECregional: Regional concentration accounting for emission density in a typical European region 
3. PEClocal: Local concentration near the emission source (within 1,000 m radius), calculated as:

1.  PEClocal = PECregional + PEClocal_source

The spatial scales are nested and specific:

• Continental: Entire Europe (3.7 million km²)
• Regional: Typical European region (200 km radius)
• Local: Immediate vicinity of source (1 km radius)

The volume of dilution is explicit and scenario-dependent: a chemical discharged into the Rhine river will have different local dilution than one discharged into a small stream in Portugal.

Output: PEClocal = 15.3 µg/L (for example)

This is an absolute concentration that can be directly compared to a toxicological threshold.

USEtox/LCA: Intrinsic Potential

When an LCA practitioner uses USEtox, they're calculating the intrinsic impact potential of a substance, independent of specific use scenarios:

Standardized input:

• Reference emission: 1 kg (pulse)

• Emission to: continental freshwater (default)

• No information about:

  - Actual tonnage

  - Specific location

  - Emission pattern

  - Receiving environment

USEtox calculates:

1. Fate Factor (FF): Residence time of the substance in freshwater, integrated over time: 

The spatial scales are designed for global applicability:

• Global: Planetary mass balance
• Continental: Generic continent (with parameterizations for specific continents)
• Urban: Outdoor urban air for human inhalation
• Indoor: Indoor air for human inhalation

The volume of dilution is implicit: built into the fate model structure but not directly accessible to the user. The continental freshwater volume is standardized, not site-specific.

Output: FF = 245 days per kg emitted (for example)

This is a time metric that will later be combined with toxicity to create a relative impact factor.

Level 2: Effect - Same Data, Different Endpoints

Both communities increasingly draw from the various ecotoxicological data sources:

• REACH registration dossiers
• EFSA pesticide evaluations
• PPDB (Pesticide Properties Database)
• Peer-reviewed literature
• USE EPA Aquatox
• RIVM data base 
• Others

The test organisms are the same:

• Algae (e.g., Raphidocelis subcapitata)
• Crustaceans (e.g., Daphnia magna)
• Fish (e.g., Pimephales promelas, Oncorhynchus mykiss)

The test guidelines are the same: OECD standardized protocols for acute and chronic toxicity.

But the endpoint derivation differs:

EUSES/Risk Assessment: PNEC (Predicted No-Effect Concentration) or HC5 (Hazardous concentration that protect 95% of the species)

The risk assessment approach aims to establish a safe threshold:

1. Identify the most sensitive species/endpoint from available data
2. Apply Assessment Factors (AF) to account for: 
3. Extrapolation from lab to field (typically AF = 10)
4. Limited species tested (AF = 10-1000 depending on data quality)
5. Acute to chronic extrapolation (AF = 10-100)
6. Uncertainty in data quality

Example calculation:

Chronic NOEC (Daphnia) = 50 µg/L (most sensitive endpoint)

Assessment Factor = 100 (limited chronic data)

PNEC = 50 / 100 = 0.5 µg/L

The PNEC represents a conservative threshold below which no effects are expected on aquatic ecosystems.

Alternatively, European guidance now allows using HC5 (Hazardous Concentration affecting 5% of species) derived from Species Sensitivity Distributions (SSD) when sufficient data are available (typically ≥10 species).

USEtox/LCA: HC20 and Effect Factor

The LCA approach aims to quantify comparative toxicity:

1. Compile all available ecotoxicity data for the substance
2. Construct a Species Sensitivity Distribution (SSD)
3. Derive HC20 (concentration affecting 20% of species)
4. Calculate an Effect Factor:

Effect Factor (EF) = 0.2 / HC20  [PAF·m³/kg]

Where PAF = Potentially Affected Fraction of species.

The Effect Factor represents the intrinsic hazard of the substance on aquatic biodiversity, using a standardized slope assumption between HC20 and HC50.

Key Difference in Toxicological Endpoints

Important note: The difference between HC5 and HC20 doesn't mean one approach is "better" or "more protective." They serve different purposes:

• HC5/PNEC: "Below this level, the ecosystem is considered safe"
• HC20: "This is a reference point for comparing relative toxicity"

Level 3: Impact/Risk - Fundamentally Different Questions

This is where the paths fully diverge.

EUSES/Risk Assessment: Risk Characterization Ratio (RCR)

The final step in risk assessment is straightforward:

RCR = PEClocal / PNEC

Example:

RCR = 15.3 µg/L / 0.5 µg/L = 30.6

Interpretation:

• RCR < 1: The predicted exposure is below the safe threshold → Safe use demonstrated ✓
• RCR > 1: The predicted exposure exceeds the safe threshold → Risk identified, risk management measures required

This is a binary, context-specific decision:

• Is this substance safe in this specific use scenario?
• For this tonnage, at this site, with these emissions?

The RCR has no transferability: if the tonnage changes, if the site changes, if the receiving environment changes, you must recalculate everything.

USEtox/LCA: Characterization Factor (CF)

USEtox combines Fate and Effect to create a Characterization Factor:

CF = FF × EF  x XF [CTUe]

Where CTUe = Comparative Toxic Unit for ecosystems

           = PAF·m³·day/kg emitted

Example:

CF = 245 days × 0.0625 PAF·m³/kg x 0.5 = 7.65 CTUe/kg

Interpretation: The CF represents the intrinsic impact potential of emitting 1 kg of this substance.

In an LCA study, the practitioner then applies this CF to the actual emission:

Impact = CF × mass emitted (from inventory)

Example: 

If LCA inventory shows 0.002 kg emitted

Impact = 7.62 CTUe/kg × 0.002 kg = 0.01 CTUe

This impact is then compared with other substances or other product scenarios:

• Product A (using substance X): Total ecotox impact = 0.0306 CTUe
• Product B (using substitute Y): Total ecotox impact = 0.0015 CTUe

Conclusion: Product B has ~20× lower ecotoxicity impact than Product A → Prefer substance Y for sustainability

The CF has universal transferability: the same CF value (15.3 CTUe/kg) is used by every LCA practitioner worldwide for any study involving this substance, regardless of:

• Geography (unless using regionalized CFs)
• Scale of emission
• Specific context

The Paradox: Same Foundation, Incomparable Outputs

Here's the critical realization for both communities:

EUSES Output: PEC = 15.3 µg/L

USEtox Output: FF = 245 days per kg emitted, CF = 7.65 CTUe/kg

These cannot be directly compared or converted!

Even though both models use SimpleBox, even though both use the same physicochemical properties, even though both could use the same toxicity data, the outputs answer fundamentally different questions:

This isn't a weakness—it's a feature. Each tool is optimized for its purpose.

A Critical Difference: Precision vs. Order of Magnitude

Beyond the different types of outputs (absolute vs. relative), there's a fundamental difference in how discriminating each approach is among substances:

EUSES/RA: High Precision Around the Decision Threshold

Risk assessment is designed to make fine distinctions near the critical RCR = 1 threshold:

Substance A: RCR = 0.85  → Safe use ✓

Substance B: RCR = 1.15  → Risk identified ✗

A difference of just 35% in RCR can change the regulatory decision. This precision is necessary because:

• Legal compliance is binary (safe or not safe)
• Risk management measures have costs
• Companies need clear guidance on what's acceptable

The PEC-based approach can discriminate among substances across a wide range of properties:

• Highly degradable substances (t½ = days)
• Moderately persistent substances (t½ = weeks)
• Persistent substances (t½ = months)
• Very persistent substances (t½ = years)

All of these can potentially yield different RCR values, depending on tonnage and scenario specifics.

RA allows fine-grained differentiation across the full spectrum of substance properties.

USEtox/LCA: Order-of-Magnitude Discrimination

In contrast, USEtox primarily differentiates substances based on large differences in environmental persistence and bioaccumulation:

Substance C: CF = 15 CTUe/kg    (low persistence, t½ water = 5 days)

Substance D: CF = 450 CTUe/kg   (moderate persistence, t½ water = 60 days)  

Substance E: CF = 8,500 CTUe/kg (high persistence, t½ water = 365 days)

USEtox excels at identifying substances that remain in the environment for extended periods—these are the substances that will dominate life cycle impact assessments because their Fate Factors (residence times) are orders of magnitude higher.

However, USEtox is less effective at discriminating between substances with:

• Similar persistence profiles
• Short to moderate degradation rates
• Differences primarily in acute toxicity rather than persistence

USEtox works by order of magnitude rather than fine gradation.

This is particularly relevant for metals and inorganics (read this post: Why Do Metals Systematically Dominate USEtox Toxicity scores?): substances that don't degrade but partition differently may show similar CFs despite different environmental behaviors.

Why This Matters for Decision-Making

This difference in sensitivity has profound implications:

Example from practice:

Imagine comparing three pesticide active substances:

Risk Assessment (EUSES):

• Pesticide X: RCR = 0.9 → Safe ✓
• Pesticide Y: RCR = 1.1 → Risk or Substance of concern ✗
• Pesticide Z: RCR = 0.8 → Safe ✓

Clear differentiation for regulatory purposes.

LCA (USEtox):

• Pesticide X: CF = 45 CTUe/kg
• Pesticide Y: CF = 52 CTUe/kg
• Pesticide Z: CF = 38 CTUe/kg

From an LCA perspective, all three are in the same ballpark (not orders of magnitude apart). The choice between them won't significantly impact the overall product's life cycle ecotoxicity footprint. However, if there were a pesticide W with CF = 2,500 CTUe/kg (highly persistent), USEtox would clearly flag this as problematic from a sustainability perspective.

The bottom line: RA provides the precision needed for compliance decisions across all substance types. USEtox provides the order-of-magnitude screening needed to identify persistence-driven impacts in life cycle thinking. Neither can replace the other.

The LCA Overclaim: Why Speed Doesn't Equal Better Protection

There's a common narrative in the LCA community that goes something like this:

"LCA evaluates thousands of substances simultaneously across the full life cycle. Risk assessment evaluates one substance at a time and takes years to cover a chemical portfolio. Therefore, LCA is more holistic and better suited for protecting the environment at a global scale."

This argument contains some truth, but it also masks critical limitations that both communities need to acknowledge honestly.

What's True: LCA's Speed Advantage

Yes, LCA can rapidly process thousands of substances in a single inventory analysis. A comprehensive product LCA might include:

• 500+ organic chemicals (solvents, additives, residues)
• 50+ metals (Cu, Zn, Cr, Ni, Pb, etc.)
• 30+ pesticide active ingredients
• Hundreds of combustion by-products (but those are rarely characterized du to lack of tox data)

All of these get characterized simultaneously using their respective CFs. This would indeed take decades to assess individually through REACH-style risk assessment.

From a systems thinking perspective, this is valuable: LCA captures life cycle stages that RA typically doesn't (e.g., raw material extraction impacts, manufacturing in distant countries, end-of-life scenarios).

What's Not True: That This Makes LCA "Better" at Protection

Here's the uncomfortable reality about ecotoxicity impact scores in LCA:

1. Metal Domination

In virtually any LCA study that includes ecotoxicity impact assessment, metals overwhelmingly dominate the score:

Typical ecotoxicity impact breakdown:

Total impact: 10,000 CTUe

Metals (Cu, Zn, Ni, Cr): 9,850 CTUe (98.5%)

All organic chemicals: 150 CTUe (1.5%)

Why? Because metals don't degrade. Their Fate Factors are essentially infinite (or limited only by burial in deep sediments). A few grams of copper released during electronics manufacturing can generate more "impact" in USEtox than kilograms of organic chemicals.

This isn't wrong scientifically—metals do persist. But it means that the ecotoxicity score is essentially a metal persistence inventory, not a comprehensive toxicity assessment.

2. Organic Chemicals Are Invisible

The consequence of metal domination is that even problematic organic substances become invisible in the aggregate score:

Imagine comparing two products:

Product A:

• Ecotoxicity score: 8,500 CTUe

• Breakdown: 8,450 CTUe from Cu/Zn, 50 CTUe from organics

• Uses a highly toxic pesticide (RCR = 5 in RA) → contributes 8 CTUe

• Uses safe solvents → contributes 42 CTUe

Product B:

• Ecotoxicity score: 9,200 CTUe

• Breakdown: 9,180 CTUe from Cu/Zn (more metal packaging), 20 CTUe from organics

• Uses no problematic pesticides

• Uses same safe solvents

LCA conclusion: Product A is "better" (lower score)

Reality: Product A uses a substance that creates severe local risks (RCR = 5), but this is completely masked by the metal-dominated score. The pesticide's 8 CTUe contribution is indistinguishable noise compared to the 8,450 CTUe from metals.

You cannot use aggregate LCA ecotoxicity scores to make informed decisions about organic chemical substitution.

3. Local Impacts Are Missing

USEtox operates at continental/global scale. It cannot assess:

• Discharge into a small stream near a manufacturing facility
• Agricultural runoff into sensitive wetlands
• Industrial accidents or spills
• Site-specific vulnerability (endangered species, drinking water sources)

These are precisely the scenarios where chemical toxicity matters most for immediate ecosystem protection. RA, with its local-scale modeling, can evaluate these scenarios. LCA fundamentally cannot.

4. Loss of Substance-Specific Information

When you aggregate thousands of substances into a single ecotoxicity score, you lose the ability to identify which specific substances are problematic:

• Is the high score due to a few very persistent organics?
• Or is it just baseline metal content?
• If I reformulate to remove substance X, will the score actually improve?
• Can I distinguish a genuinely safer alternative from a slightly different metal alloy?

The answer is often no—the aggregate score doesn't provide actionable guidance for chemical substitution decisions involving organic chemicals.

The Consequence: LCA Creates an Illusion of Comprehensive Protection

The danger of LCA's "thousands of substances evaluated" claim is that it creates a false sense of comprehensive chemical safety assessment:

• Stakeholders see "ecotoxicity impact assessed" ✓
• Companies report "we evaluated all chemicals in our product" ✓
• Consumers believe "LCA means this product is environmentally safe" ✓

But none of this is true if the dominant impact is metal persistence and the organic chemicals of actual concern are invisible in the score.

This isn't to say LCA is useless—it's to say that LCA ecotoxicity scores answer a different question than "is this product's chemical composition safe?"

What LCA ecotoxicity actually tells you:

• ✅ Which products/processes have very high persistence-driven impacts
• ✅ Whether metal content is a significant environmental consideration
• ✅ Order-of-magnitude differences in persistence between alternatives

What LCA ecotoxicity does not tell you:

• ❌ Whether specific organic chemicals create local risks
• ❌ Whether a product meets chemical safety requirements
• ❌ Which of two similar pesticides is safer to use
• ❌ Whether reformulation eliminated a problematic substance

What Risk Assessment Provides That LCA Cannot

Despite being "slow" and "one substance at a time," RA delivers something LCA cannot: granular, substance-specific protection.

RA advantages:

1. Substance-level resolution: Each chemical evaluated independently—no masking by metals
2. Local impact assessment: PECs calculated for actual receiving environments, protecting vulnerable ecosystems
3. Actionable differentiation: Can distinguish safe (RCR = 0.8) from unsafe (RCR = 1.2) substances with similar persistence
4. Risk management integration: Defines safe use conditions, required controls, exposure limits
5. Regulatory accountability: Companies must demonstrate safe use or face restrictions

Yes, this takes time. Yes, it's resource-intensive. But it actually protects ecosystems in a way that aggregate LCA scores often don't.

The Reconciliation: Both Are Necessary, Neither Is "Better"

The key message for both communities:

For LCA practitioners:

• Stop claiming LCA is "better" or "more comprehensive" than RA for chemical safety
• Acknowledge that your ecotoxicity scores are dominated by metal persistence and largely insensitive to organic chemical choices
• Use LCA for what it does well: identifying persistence hotspots and system-level impacts
• Don't use LCA ecotoxicity scores alone for chemical substitution decisions involving organics

For risk assessors:

• Acknowledge that RA's substance-by-substance approach misses cumulative effects and life cycle burdens
• Recognize that LCA can identify persistent substances that might pass local RA but accumulate globally
• Understand that LCA's speed enables screening-level assessments useful for early R&D
• Don't dismiss LCA just because it operates at a different scale

The honest claim for LCA:

• ❌ "LCA protects the planet better than RA because it evaluates thousands of substances"
• ✅ "LCA complements RA by identifying persistence-driven impacts and system-level burdens that substance-by-substance RA might miss"

The honest claim for RA:

• ❌ "RA is sufficient; LCA adds no value"
• ✅ "RA protects against immediate risks through precise substance assessment; LCA reveals long-term persistence concerns and cumulative system impacts"Engage with this topic

What's your experience working across the LCA and risk assessment communities? Have you encountered challenges in communicating between these disciplines? Share your thoughts in the comments or connect with me on LinkedIn to continue the discussion.