LCA Technology cheat sheet

ISO 14040 defines the four-phase framework: goal and scope definition, inventory analysis (LCI), impact assessment (LCIA), and interpretation; phases are iterative, not strictly linear.

ISO 14044 supplies the detailed requirements: goal and scope must state intended application, reasons for the study, audience, commissioner, and whether a comparative assertion disclosed to the public is intended.

Comparative assertions disclosed to the public trigger mandatory critical review under ISO 14044 and require a panel review including interested parties.

The functional unit is the quantified performance of the product system used as the reference unit; the reference flow is the amount of product needed to deliver that function, and all inputs/outputs are scaled to it.

System boundary defines which unit processes are included; ISO 14044 requires decisions on capital goods, infrastructure, and cut-off criteria to be documented and justified.

Allocation follows the ISO 14044 hierarchy: (1) avoid allocation by subdivision or system expansion; (2) if unavoidable, allocate via physical relationships; (3) otherwise by other relationships such as economic value.

A cut-off or exclusion criterion (e.g., mass or energy share, environmental significance) must be stated with its threshold; excluded flows cannot be assumed negligible without justification.

The product system is a set of unit processes connected by flows; elementary flows cross the boundary to the environment, while product flows stay inside the system.

Time, geography, and technology coverage of the system are part of the scope and must be reported as they constrain data selection and comparability.

Foreground data are collected by the study practitioner for the specific product system; background data come from third-party databases and represent upstream supply chains.

Aggregated datasets collapse many unit processes into one; their internal structure cannot be modified, so allocation and boundary choices embedded in them are inherited by the user.

Common background sources include ecoinvent, GaBi/Sphera, Agribalyse (agriculture), and national LCI databases; each has its own geography, reference year, and system model.

ecoinvent system models (cut-off, APOS, consequential) differ in how by-products and recyclable materials are handled; mixing models within one study distorts results.

Data quality indicators typically cover temporal, geographical, and technological representativeness, completeness, precision/uncertainty, and methodological appropriateness.

The pedigree matrix converts qualitative data quality scores into uncertainty factors that are combined with basic uncertainty to yield a lognormal distribution.

Lognormal uncertainty is described by geometric mean (the value) and geometric standard deviation (GSD); GSD 2 means a 95% range of roughly 0.25x to 4x the value.

Elementary flow names must match the reference substance of the LCIA method (e.g., 'Carbon dioxide, fossil' to the global warming substance list); naming mismatches cause silent zero impacts.

Mandatory LCIA elements: selection of impact categories, classification (assigning elementary flows to categories), and characterization (applying characterization factors). Optional: normalization, grouping, weighting.

Midpoint indicators sit close to emissions (e.g., kg CO2-eq, kg N-eq) and are more certain; endpoint indicators model damage to human health, ecosystems, and resources and add more uncertainty.

Characterized result = sum over flows of (elementary flow quantity x characterization factor); CFs are method- and sometimes context-specific (e.g., spatialized or time-horizon dependent).

Global warming characterization factors depend on the time horizon: GWP100 is standard; GWP20 and GWP500 differ substantially, so the horizon must be reported.

Method families: CML and TRACI are midpoint; ReCiPe and IMPACT World+ offer both midpoint and endpoint; EF (Environmental Footprint) provides a recommended midpoint set with normalization and weighting.

Normalization divides characterized results by a reference system (e.g., per-capita emissions in a region/year); weighting then aggregates categories using value-based factors and should be reported as optional and subjective.

ReCiPe endpoint results use a cultural perspective (individualist, hierarchist, egalitarian) that changes time horizon and damage modeling assumptions.

Method selection must be consistent with the goal; comparative studies require the same method, version, and spatial resolution across all systems compared.

openLCA, SimaPro, GaBi/Sphera, and Umberto are common modeling environments; each stores flows, processes, and impact methods in its own database schema.

Parameter sets allow one model to represent many variants; local parameters apply within one process, global parameters apply across the database, and formulas can reference both.

Scenario analysis is implemented by varying parameter values or swapping process links; results should be traceable to the specific parameter set used.

Product systems in openLCA are built from linked processes and can be exported as JSON-LD; the product system stores provider links and quantitative reference.

Version control: database snapshots, exported archives, and documented model versions are needed to reproduce results; software and database version numbers belong in the report.

Interoperability routes include ILCD zip packages, ecoSpold2, and openLCA JSON-LD; conversions may lose metadata such as uncertainty or method mapping.

Allocation settings and system-model choices are properties of the database/model, not of the impact method; changing them changes inventory results, not characterization.

Unit consistency across flows is enforced by flow properties; mismatched units or missing conversion factors are a frequent source of modeling errors.

ISO 14044 interpretation steps: identify significant issues, evaluate them via completeness, sensitivity, and consistency checks, then draw conclusions, limitations, and recommendations.

Contribution analysis ranks processes, flows, or impact categories by their share of total impact; it identifies hotspots but not the cause of uncertainty.

Sensitivity analysis varies one parameter or methodological choice (e.g., allocation rule, GWP horizon) and reports the change in results; it tests robustness, not statistical spread.

Scenario analysis compares alternative plausible futures (e.g., energy mixes, end-of-life routes) rather than parameter ranges around one baseline.

Uncertainty propagation (Monte Carlo simulation) samples input distributions to produce a distribution of results; the number of iterations determines the stability of the output statistics.

Pedigree-based uncertainty and parameter uncertainty can be combined; correlated parameters must be sampled jointly or results overstate spread.

Consistency check verifies that the same method, boundary, and data quality rules were applied uniformly across compared systems.

Completeness check confirms that all relevant flows and processes within the boundary are accounted for; documented gaps must be carried into the conclusions.

Reporting must state results, limitations, and the effect of methodological choices so a reader can judge whether conclusions hold beyond the studied case.

ILCD format (European Commission) is an XML-based dataset and process format used for compliance with the International Reference Life Cycle Data System; ILCD packages bundle datasets with method and flow properties.

ecoinvent ecoSpold2 is an XML schema for activities, exchanges, and master data; it supports uncertainty fields and is the native export of the ecoinvent database.

openLCA JSON-LD is a JSON schema using linked-data identifiers for flows, processes, and impact methods; it is the native import/export format of openLCA.

EPD formats are product-category-specific (e.g., EN 15804 for construction products, EN 50693 for electrical/electronic); PCRs define the rules, and EPD data are often exchanged as ILCD or via the ECO Platform/International EPD System.

Flow and unit harmonization is the main interoperability failure point: the same substance may appear under different UUIDs, names, or compartments across databases.

Data exchange workflows typically run: export source dataset, map flows to target nomenclature, validate units and reference flows, then import and reconcile impact method mapping.

Digital product passports and PEF-aligned data rely on structured, machine-readable formats; unstructured PDF EPDs lose machine readability and cannot be aggregated automatically.

Version and UUID preservation matter for traceability; re-importing converted data can break links to impact methods and uncertainty records.

ISO 14044 defines critical review as a process ensuring consistency with the standard; it can be by internal expert, external expert, or a panel of interested parties.

Panel review is required for comparative assertions disclosed to the public and must include at least one independent expert plus representatives of interested parties; the review statement is published with the study.

Review scope covers methodology, data quality, interpretation, and reporting; the reviewer's comments and how they were addressed should be documented.

Verification of an EPD or carbon footprint is performed by an approved verifier against the applicable PCR and program instructions, not against ISO 14044 alone.

Reproducibility requires documented data sources, software and database versions, parameter values, allocation decisions, and method versions.

Transparency means a reader can trace results back to inventory data and characterization factors; aggregated-only reporting limits transparency.

Internal QA checks include mass and energy balance checks, unit consistency, completeness against the boundary, and plausibility comparison with literature ranges.

A review statement must state the review type, reviewer independence, and whether the study meets the requirements of ISO 14040 and 14044.

An EPD reports environmental performance of a product against a declared unit (or functional unit) following a PCR; EN 15804 defines modules A1-A3 (product stage), A4-A5, B1-B7 (use), C1-C4 (end of life), and D (benefits beyond the system).

Product carbon footprint (PCF) is a single-category LCA restricted to greenhouse gases; ISO 14067 specifies requirements, and PAS 2050 is an earlier specification.

Science-based targets and GHG Protocol Scope 3 accounting use LCA-derived emission factors; Scope 3 category 1 (purchased goods) often dominates and relies on supplier-specific or spend-based data.

Circularity metrics (e.g., Material Circularity Indicator, recycled content, end-of-life recovery rate) are complementary to LCA and do not replace impact assessment.

Digital product passports (EU ESPR context) require structured product data, including environmental and material information, that can be machine-read and exchanged across the value chain.

Automation via APIs and scripted model generation (e.g., openLCA IPC/API, Python) enables large-scale scenario modeling but requires versioned parameter inputs to remain reproducible.

PEF (Product Environmental Footprint) provides a harmonized method with prescribed impact categories, data quality requirements, and EF datasets; its use is increasingly required in EU policy contexts.

Comparative claims based on LCA must be backed by a critical review and consistent functional units; marketing claims without review risk greenwashing findings.