TNFD Mitigation Hierarchy: Avoid, Minimise, Restore and Offset Explained

Without sequencing, an organisation can move too quickly to restoration or offsetting because those actions are easier to communicate than changing the project, product or sourcing decision that creates the impact. The hierarchy forces a more difficult question first: can the impact be prevented?

The sequence also improves transparency. It separates gross negative impacts from management actions, identifies what remains after each step and makes trade-offs visible. This is important for TNFD risk and impact management, metrics, targets, transition plans and public claims.

Figure 1. The hierarchy should be evidenced as a sequence; later steps do not erase the need to test earlier options.

Avoidance changes the decision so that the negative impact does not occur. It can include selecting a different site or route, protecting set-asides, changing project design or timing, eliminating a harmful input, excluding high-risk suppliers, choosing a lower-impact material, declining finance for an unacceptable activity or redesigning a product.

A strong avoidance record shows the alternatives considered, criteria used, ecological and social evidence, commercial and technical constraints, who decided and whether the avoided impact was displaced elsewhere. “No feasible alternative” should not be a generic statement; the organisation should define feasibility and retain the analysis.

Avoidance can also be embedded in policy: no-conversion commitments, exclusion lists, no-go zones, design standards, sourcing requirements and investment screens. The control should test implementation rather than policy existence alone.

Minimisation reduces magnitude, duration, extent, frequency or likelihood after avoidance options have been applied. Examples include smaller footprints, reduced abstraction, pollution prevention, seasonal operating restrictions, wildlife corridors, better process efficiency, lower-impact harvesting, bycatch controls, transport changes and supplier corrective action.

The evidence should connect the control to the pressure it is expected to reduce. Measure both implementation and performance. Installing treatment equipment is an activity; reduced pollutant load is a pressure outcome. A supplier code is an action; verified changes in sourcing or production are performance evidence.

Minimisation should include control effectiveness and residual impact. If a control fails or uncertainty is high, the action plan should include monitoring, adaptive management and escalation.

Restoration seeks to repair damage or recover ecological functions, while regeneration is often used more broadly for actions that enable nature to recover and improve. The exact terminology and method should follow the applicable context and guidance.

A restoration plan should identify:

the affected location and ecosystem;

the baseline or reference condition;

the impact and residual pressure;

the restoration objective and ecological indicators;

methods, timing and responsible experts;

maintenance and monitoring period;

risks, dependencies and climate interactions;

community and rights-holder engagement;

criteria for success, failure and corrective action; and

long-term governance and funding.

Hectares treated or trees planted are not sufficient on their own. They are activity metrics. The organisation should monitor survival, ecosystem condition, connectivity, species response, hydrology or other relevant outcomes.

An offset is intended to compensate for significant residual adverse biodiversity impacts after appropriate avoidance, minimisation and restoration. It is not a general substitute for operational improvement and should not be assumed to be available or credible for every impact.

A robust residual-impact decision considers:

whether all reasonable avoidance options were genuinely assessed;

the effectiveness and uncertainty of minimisation and restoration;

the biodiversity values and ecosystem functions affected;

whether losses are irreplaceable, highly threatened or subject to no-go rules;

spatial and temporal equivalence;

additionality beyond existing obligations;

permanence and long-term management;

leakage or displacement of impacts;

implementation, governance and financial security;

affected communities and rights holders;

monitoring and independent expertise; and

the risk that the offset will not deliver the intended outcome.

IFC Performance Standard 6 uses “like-for-like or better” and measurable conservation outcomes in its project-finance context. Those concepts should not be copied into a corporate claim without testing applicability and evidence.

SBTN's AR3T framework is based on the mitigation hierarchy but uses a broader corporate-action framing:

Avoid and Reduce pressures on nature loss;

Regenerate and Restore so nature can recover; and

Transform underlying systems and drivers.

Transformation can include changing product portfolios, market rules, finance, incentives, infrastructure or sector practices. It is not a reason to bypass avoidance at a specific location. A company may need both: avoid a direct impact and work collectively to transform the system that creates repeated pressure.

Use the terminology accurately in disclosures. “We apply AR3T” should be supported by actions and evidence across the stated boundary, not merely by one restoration project.

Gate 1 - is the impact material or potentially severe?

If no, retain the evidence and monitoring conclusion. If yes or uncertain, proceed with specialist assessment and governance.

Gate 2 - can the impact be avoided?

Test location, design, timing, technology, material, supplier, customer and business-model alternatives. If an option is rejected, document the reason and approval.

Gate 3 - what can be minimised?

Define controls, expected pressure reduction, performance indicators and residual impact after control.

Gate 4 - what restoration or regeneration is possible?

Set the ecological objective, reference condition, method, timeframe, monitoring and failure response.

Gate 5 - does a significant residual impact remain?

If no, monitor and disclose the outcome. If yes, assess whether compensation is permissible and credible. Some impacts may be effectively non-offsettable or subject to legal or policy constraints.

Gate 6 - can an offset meet the required quality and rights safeguards?

If not, reconsider the activity, design or acceptance of residual impact. If yes, approve it separately, retain gross impact disclosure and monitor delivery over the necessary period.

Figure 2. The process may return to the underlying decision when residual impacts are unacceptable or cannot be credibly compensated.

Keep separate records for:

the gross impact before mitigation;

the impact avoided through a changed decision;

the reduction achieved through controls;

restoration or regeneration outcomes;

the residual impact remaining; and

any offset or compensation result.

This separation avoids misleading netting. A corporate dashboard may summarise the sequence, but the methodology should not subtract hectares restored in one ecosystem from conversion in another unless a credible, approved framework supports the comparison.

Claims such as “no net loss”, “net gain”, “nature positive”, “fully mitigated” or “biodiversity neutral” are high risk. They require defined boundary, metric, baseline, counterfactual, location, timeframe, equivalence, uncertainty, safeguards and governance. Even a technically designed project offset does not automatically support a company-wide claim.

An action that benefits one metric can create another impact. Water efficiency may increase energy use; a new material may shift land or pollution pressure; restoration may restrict community access; supplier exclusion may move production to a less visible region.

The mitigation decision record should therefore include:

nature trade-offs across realms and locations;

climate and social interactions;

effects on Indigenous Peoples, Local Communities and affected stakeholders;

impact displacement and leakage;

cumulative and indirect impacts;

uncertainty and reversibility; and

grievance, remedy and adaptive-management arrangements.

Engagement should inform the decision and safeguards; it should not be used as a substitute for ecological assessment or rights obligations.

A company proposes a logistics hub close to a wetland. The initial design would remove habitat and alter drainage. The first response is not to purchase restoration credits. The project team tests alternative sites and layouts, moves the main footprint away from the highest-value area and protects a buffer. It then reduces the remaining impact through drainage design, construction timing, lighting controls and access restrictions.

Disturbed areas are restored against a defined reference condition and monitored for hydrology, vegetation and species use. A residual fragmentation impact remains. The company assesses whether it is acceptable and whether an offset is legally and ecologically appropriate. The offset decision is reviewed separately, includes long-term management and community safeguards, and is not deducted from the project's gross footprint metric in the TNFD report.

Illustrative scenario only; project decisions require site-specific legal, ecological and social assessment.

Starting with offsets because the project is already approved. Reopen alternatives and evidence whether avoidance remains possible.

Calling compliance controls minimisation without performance data. Link controls to measured or evidenced pressure reduction.

Counting trees planted as restoration success. Use ecosystem- and location-relevant outcome indicators over a suitable period.

Treating residual impact as the original impact after a policy statement. Quantify or assess the impact after implemented and effective controls.

Assuming every impact is offsettable. Test irreplaceability, sensitivity, legal limits, equivalence, time lags and failure risk.

Netting unrelated locations and ecosystems. Preserve gross and location-specific information; avoid unsupported composite balances.

Using AR3T language without transformation or outcome evidence. Map each stated action to the actual framework step and boundary.

Making broad nature-positive claims from one project. Restrict claims to the evidenced outcome and obtain technical and legal review.

The accompanying TNFD Metrics, Targets and Mitigation Register includes a mitigation-action register with hierarchy step, avoidance rationale, residual-impact fields, offset-quality evidence and review status.

SRC-01 - TNFD Recommendations v1.0: https://tnfd.global/wp-content/uploads/2023/08/Recommendations_of_the_Taskforce_on_Nature-related_Financial_Disclosures_September_2023.pdf

SRC-12 - TNFD glossary: https://tnfd.global/publication/tnfd-glossary/

SRC-10 - SBTN Act / AR3T: https://sciencebasedtargetsnetwork.org/companies/take-action/act/

SRC-13 - IFC Performance Standard 6: https://www.ifc.org/content/dam/ifc/doc/2010/2012-ifc-performance-standard-6-en.pdf

SRC-14 - IFC Guidance Note 6: https://www.ifc.org/content/dam/ifc/doc/2010/2012-ifc-guidance-note-6-en.pdf