Savanna Fire Management

Tropical savannas account for around 30% of terrestrial net primary productivity and about 70% of the area burned worldwide each year¹, equivalent to about 3.3 petagrams of carbon annually, or nearly one third of emissions from fossil fuel combustion². Most savanna burning, and most of its non-CO₂ Greenhouse Gas (GHG) emissions, occurs late in the dry season, when cured fuels, high temperatures, low humidity and strong winds produce intense, fast-spreading fires. A global assessment places the baseline savanna fire source at 749 teragrams of CO₂-equivalent per year from carbon monoxide, methane and nitrous oxide, with 70% of emissions occurring in the late dry season³.

Fires lit early in the dry season burn less completely and uniformly than late dry season fires. Field measurements in southern African miombo woodlands, for example, show that early season fires consume 6 to 13 % less fine fuel than late season fires. They also burn 65 to 68% of the area within the fire scar, compared with 82 to 96% for late season fires. Overall, early season emissions per unit area are estimated at roughly 34 to 56% of late season levels⁴. Nonetheless, the largest source of abatement is the reduction in total area burned that early, patchy burning produces, rather than the change of season in itself.

Sustained fire management also affects the carbon held in woody biomass. Long-term Australian trials find that early dry season fire regimes accumulate more woody carbon than late dry season regimes, and northern Australia's savannas shifted from a net source of approximately 19 megatonnes of CO₂e per year over 1990 to 2015 to a net sink of 3.5 megatonnes per year over 2016 to 2022 as savanna fire management was adopted across the region^5,6.

This Protocol provides the requirements and procedures for quantifying the net greenhouse gas impact of savanna fire management. Savanna fire management shifts burning from the late dry season to the early dry season and typically reduces the total area burned, lowering fire emissions and allowing woody biomass to accumulate. This Protocol credits two outcomes: the reduction of methane (CH₄) and nitrous oxide (N₂O) emissions from fire, and the removal of carbon dioxide (CO₂) from the atmosphere into woody biomass.

This Protocol is developed to adhere to the requirements of ISO 14064-2:2019. It ensures that:

• emission reductions (CH₄ and N₂O) and woody-biomass removals (CO₂) are quantified separately and credited without double counting;
• fire emissions are quantified consistently for both the Project and counterfactual scenarios, stratified by season and Vegetation Fuel Type;
• the reductions baseline is the project area's own historical mean fire emissions and the removals baseline is a dynamic matched control, and both are conservative;
• credited reductions and removals are additional, independently verified, and quantified at a conservative lower bound;
• woody-biomass removals are durable, with Reversals compensated from a removals Buffer Pool;
• fire management is community-led, respects Indigenous and traditional fire knowledge, and safeguards biodiversity; and
• potential leakage is quantified.

Throughout this Protocol, "must" indicates a requirement, "should" indicates a recommendation, "may" indicates a permitted option, and "will" indicates a statement of fact or future certainty.

This Protocol relies on and is intended to be compliant with the following standards:

• The Isometric Standard
• ISO 14064-2:2019, Greenhouse Gases, Part 2: Specification with guidance at the project level for quantification, monitoring and reporting of greenhouse gas emission reductions or removal enhancements

Additional reference standards that inform the requirements and practices in this Protocol include:

• ISO 14064-3:2019, Greenhouse Gases, Part 3: Specification with guidance for the verification and validation of greenhouse gas statements
• ISO 14040:2006, Environmental Management, Life Cycle Assessment, Principles and Framework
• 2006 IPCC Guidelines for National Greenhouse Gas Inventories, Volume 4, Chapter 6 (Grassland)⁷

Methodologies that were assessed during the development of this Protocol include:

• Carbon Farming Initiative (Savanna Fire Management,) Determination, Australian Carbon Credit Unit (ACCU) scheme, 2015
• Carbon Farming Initiative (Savanna Fire Management, Sequestration and Emissions Avoidance) Determination, Australian Carbon Credit Unit (ACCU) scheme, 2026
• Savanna Fire Management, consultation draft (CN0135), Verra, 18 February 2026
• Savanna Fire Management methodology submission, Paris Agreement Article 6.4 Mechanism, 2025

This Protocol is built on the current state of published science on savanna fire management and the long-term monitoring of savanna carbon. The field is developing quickly, and several requirements are expected to tighten as evidence accumulates. Anticipated improvements include:

• Replacing the early and late dry season surrogate with remotely sensed fire severity¹;
• Including emission factors for carbon monoxide (CO);
• Broadening the eligible regions and Vegetation Fuel Types as region-specific emission factors and fuel parameters are validated;
• Validating fuel accumulation with field measurements;
• Using satellite measures of fuel curing;
• Incorporating the effects of varying levels of herbivory⁸.

This Protocol will be reviewed when an update to published science, government policy or legal requirements would affect net greenhouse gas quantification or the monitoring requirements set out here, and at a minimum every two years.

Non-woody grassland (<10% woody canopy cover) is excluded because low-intensity early dry season fire there can promote woody thickening and encroachment. Seasonally waterlogged and riparian grasslands are excluded because their fuels remain uncured late into the dry season.

The Project is credited on the measured reduction in fire emissions and on the measured increase in woody-biomass carbon, not on the share of fire that occurs in the early dry season. Fire emissions are quantified separately for the early and late dry season, within each Vegetation Fuel Type (VFT) and fuel class, on both the baseline and the project side. A Project that shifts fire to the early dry season but burns more total area, or burns more often, therefore earns proportionally fewer Credits.

This Protocol does not require an exclusively early dry season fire regime. Prescribing only early dry season burning in mesic savannas prone to woody thickening can undermine the long-term emissions objective and create biodiversity trade-offs⁹.

In each Reporting Period, burned area is classified against the Default Seasonal Cutoff Date for the project’s region (Appendix A, Table A1b), which is known to the Project in advance of each burning season, and is reclassified to the late dry season where that year’s night-time fire behavior in the Seasonal Observation Area shows the transition occurred materially earlier. A Project may adopt an earlier date, but not a later one.

Projects must use the same Seasonal Observation Area, rainfall dataset and active-fire products for the historical record and the project period.

This Protocol applies to fire-prone woody savannas, typically with woody canopy cover >10%⁴.

Emission factors and fuel parameters are region-specific and cannot be generalized across regions of the world or other vegetation types^3,10,11. Further regions, and further Vegetation Fuel Types, become eligible once their parameters are validated by field studies and in consultation with Isometric.

A Vegetation Fuel Type is a mapped class of structural formation, canopy cover, and grass type, and each carries its own fuel-load accumulation, combustion-completeness, patchiness and emission-factor parameters. Vegetation Fuel Types are the spatial unit of emissions accounting, and thus improper stratification biases every emission estimate. A Project may only enroll land in the Vegetation Fuel Types that Appendix A defines for its region and that meet the eligibility thresholds set there.

The baseline scenario is the continuation of the project area's pre-project fire regime, characterized from the project area's own historical fire record.

The baseline is fixed before project activities begin and is not reassessed after the project starts.

Distinguishing early- from late-dry-season fire is central to the emissions reductions this Protocol credits. All burned areas are classified by season and that classification drives the seasonal emission contrast.

Projects must use the same rainfall dataset for the historical record and the Crediting Period.

Default Seasonal Cutoff Date. The Default Seasonal Cutoff Date is the fixed early/late dry season boundary given for the project’s region in Appendix A, Table A1b. It is known before each burning season and is used to classify burned area in every Reporting Period during project years, subject to the adjustment for an earlier seasonal transition date. In the historical record it is used only for years before 2012 and for years in which no seasonal transition date can be determined. A Project may adopt an earlier date, fixed before the Crediting Period begins, but not a later one.

Reporting Periods. During the project years, burned area in each Reporting Period is classified against the Default Seasonal Cutoff Date. Projects must reclassify to the late dry season all burned area observed between the transition date and the Default Seasonal Cutoff Date where that year's transition date falls more than ten days earlier than the Default Seasonal Cutoff Date. The cutoff is never moved later. Where no transition date can be determined, the Default Seasonal Cutoff Date applies without adjustment.

Burn dating. Projects must date burned area to the narrowest interval the imagery allows and report the dating precision. Where the burn date of a pixel cannot be resolved to one side of the applicable cutoff, it is classified as late dry season in the Reporting Period and as early dry season in the historical record.

Fuel curing. In a Reporting Period, burned area observed earlier than 28 days after the start of that year's annual dry season, or after the end of the annual dry season, is classified as an uncured burn. Uncured burns are quantified with the late dry season parameters, and their emissions are multiplied by two (Equation 2). The grass layer of eligible woodland can burn before it has cured, and burning uncured grass lowers combustion efficiency and raises methane yield^17,18,19. The multiplier is a conservative allowance for this and applies to the project side only.

Off-season fires. Projects must count any fires detected outside the dry season. In a Reporting Period they are uncured burns, as above. In the historical record, burned area observed outside the annual dry season is classified as early dry season. Future versions will seek to use remotely sensed fire severity (see Section 3.0).

This Protocol credits two outcomes with different temporal profiles. Emission reductions are an annual flux that continues for as long as the Project sustains the managed fire regime. Woody-biomass removals are a stock that accumulates while the managed fire regime allows woody biomass to grow, then levels off as the biomass reaches a new equilibrium. The crediting structure reflects this difference.

Projects must provide the following to evidence the length of the Project Commitment Period:

Project Termination. Abandonment or failure to perform project activities at any point in the Project Commitment Period, where not remedied by the Project Proponent or a successor within a reasonable period, may result in project failure. Project failure is treated as a Reversal of all Removal Certificates issued under the Project. Reduction Certificates already issued are not affected, and any outstanding reductions debt is made good as set out in Section 10.3.4.

A Project that has not previously undertaken a fire management climate intervention on the project area begins with an initial Crediting Period of up to 25 years. During the initial Crediting Period, the Project credits both pools: the reduction in CH₄ and N₂O fire emissions, and the removal of CO₂ into woody biomass. Removals are credited where the Project measures an increase in woody-biomass stock in accordance with Section 9.0. A Project without measurable, eligible woody-biomass accumulation credits emission reductions only.

Projects must indicate the last Reporting Period of the removals Crediting Period to be submitted for Verification.

Failure to quantify removals at a Verification within 5 years of the previous Reporting Period in which removals were quantified, or to request an extension, will conclude the removals Crediting Period and shift the removals account into the Ongoing Monitoring Period.

The Ongoing Monitoring Period begins at the end of the removals Crediting Period and runs for the length committed at PDD submission. Every Project crediting removals has an Ongoing Monitoring Period, whose length sets the durability of the Removal Certificates (see Section 5.4).

No Removal Certificates are issued during the Ongoing Monitoring Period. Monitoring for Reversals is conducted by Isometric throughout the Ongoing Monitoring Period, as described in Section 10.4, and Reversals are compensated by the removals Buffer Pool (see Section 10.3).

A loss of credited woody-biomass stock, whether from severe fire, drought or other disturbance, is a Reversal and is compensated from the removals Buffer Pool, in accordance with the Isometric Standard.

The durability of Removal Certificates is set to the length of the Ongoing Monitoring Period. The Project Commitment Period may be extended later only by extending the removals Crediting Period; renewal of the reductions Crediting Period does not extend it (see Section 5.5). Projects must not change the Ongoing Monitoring Period from the length committed at PDD submission.

For example, a Project with an initial Crediting Period of 20 years that commits to an Ongoing Monitoring Period of 30 years has a Project Commitment Period of 50 years, and its Removal Certificates have a durability of 30 years. If it later extends the removals Crediting Period by 5 years, the Project Commitment Period becomes 55 years and the Ongoing Monitoring Period remains 30 years. Its reductions Crediting Period may be renewed during the Ongoing Monitoring Period, but no renewal may run beyond year 55.

The reductions pool may be renewed and continue to issue emission reductions independently of the removals account, whether the removals account is in an extended Crediting Period or the Ongoing Monitoring Period, subject to the renewal conditions of the Isometric Standard. Because the reduction in fire emissions is an ongoing flux rather than a stored stock, it does not carry a permanence obligation, and renewal allows it to continue to be credited for as long as the Project sustains the managed fire regime and remains additional. For a Project crediting removals, a renewal may run in parallel with the Ongoing Monitoring Period but must not run beyond the end of the Project Commitment Period, and does not extend it (see Section 5.1).

The baseline is not reassessed on renewal (see Section 8.0) given that the project area conditions have been altered by the previous Crediting Period. Projects must confirm at each renewal that the baseline scenario still applies, that they still meet the eligibility conditions of Section 4.0, and that they still hold the legal right to implement prescribed burning, including any permit or exemption required under a fire-exclusion law (see Section 7.3).

The legal right to implement and legal compliance are also confirmed at each verification.

The Reporting Period is the interval over which reductions and removals are quantified and verified. The fire record is monitored annually, because the annual fire cycle is the unit of analysis and fire emissions are dominated by a small number of high-fire years. The minimum Reporting Period length is one year. Certificates are issued for each verified Reporting Period.

Woody biomass accumulates over longer time scales than the annual fire cycle, and its change over a single year is often smaller than the measurement error. Projects must quantify reductions for every Reporting Period. Projects may choose the frequency at which they quantify removals for Certificate issuance, by quantifying removals only at the end of the Reporting Periods they select.

In any Reporting Period in which the project area's fire emissions exceed the baseline, the result is a net debit (net emissions rather than net reductions), which is drawn from the Reductions Escrow (see Section 10.3.4). Because the baseline is constructed to be unbiased over the pre-project record, annual over-estimates and under-estimates net out across the Crediting Period, which is why Certificates may be issued annually. Crediting is not conditional on the Project reaching a steady fire regime.

This section sets out the safeguards specific to savanna fire management. They operate on top of, not instead of, the requirements that every Isometric project inherits from the Isometric Standard, and adherence to them, with independent verification, is a condition of crediting.

Because savanna fire management changes a practice that communities depend on and carry out themselves, the social safeguards below are central to the integrity of the Project rather than peripheral to it: the people who set fire in and around the project area are its fire managers, and excluding or displacing them is at once an ethical failure, an effectiveness failure, and a driver of leakage⁹.

Projects must develop and implement a Stakeholder Engagement Plan in accordance with the Stakeholder Input Process of the Isometric Standard.

The VVB may interview or survey stakeholders to confirm this.

Integrating local communities who utilize the land for their livelihoods into project design and planning works as an upstream control on activity-shifting leakage (see Section 8.3). Communities that manage and benefit from the fire regime are not incentivized to relocate fire practices beyond the project boundaries.

Projects must:

Early dry season burning pursued without regard to fire-regime diversity can homogenize habitat and harm fire-sensitive species²¹.

Projects must set and commit to biodiversity-protective fire-regime targets that have a scientific basis and are appropriate to the Project’s region.

The commitment is monitored from the fire-regime data the Project already produces (season, frequency, time since last burnt, and patch size).

Projects must remediate any committed value not met in a Reporting Period, in accordance with the Isometric Standard; failure to remediate may result in cessation of Certificate issuance.

The following topics are governed by the Isometric Standard, which applies to all Isometric Protocols. They are summarized here only as they apply to savanna fire management; the Standard is authoritative.

Projects must prepare a Project Design Document (PDD), as required by the Isometric Standard, that includes:

• the Project Boundary, the Vegetation Fuel Type map and its accuracy assessment, and any ineligible vegetation excluded from quantification;

• the pre-project fire record, the fixed baseline, the Seasonal Observation Area, and the removals Donor Zone;

• a fire-management plan, including the role of Indigenous and traditional fire knowledge and the community-led management arrangements;

• the biodiversity fire-regime commitment and its scientific basis;

• evidence of the legal right to conduct prescribed burning and rights to reduction and removal claims;

• the monitoring plan for fire, woody biomass, leakage and reversals.

Projects must be validated, and their net reductions and removals verified, by an independent Validation and Verification Body (VVB), consistent with this Protocol and the Isometric Standard.

The VVB must consider the following requisite components:

• Verify that the Project meets the Applicability conditions described in Section 4.0.
• Verify that the Environmental and Social Safeguards described in Section 6.0 are met.
• Verify that the System Boundary and Leakage assessment adhere to the requirements of Section 8.0.
• Verify that the quantification approach and monitoring plan adhere to the requirements of Section 9.0.
• Verify that the conditions for ensuring durability and monitoring for Reversals in Section 10 are met.
• Verify that the Project is compliant with the requirements of the Isometric Standard.

As part of this evaluation, the VVB must also review the characterization and quantification of the individual uncertainty sources that contribute to the net reduction and net removal. Fire is monitored annually (see Section 5.0); the cadence of formal third-party verification is set by the Isometric Standard.

The threshold for Materiality, across the totality of all omissions, errors and misstatements, is 5% of the net reduction or removal, in accordance with the Isometric Standard.

Verifiers should also verify the documentation of uncertainty of the GHG Statement, as required by the Isometric Standard. Qualitative Materiality issues may also be identified and documented, such as:

• control issues that erode the verifier's confidence in the reported data;
• poorly managed documented information;
• difficulty in locating requested information; and
• noncompliance with regulations indirectly related to GHG emissions, reductions or removals.

The VVB must conduct site visits to the project area in accordance with ISO 14064-3, 6.1.4.2, and the Isometric Standard, including at a minimum one during Validation and one during the initial Verification, at which the validator observes fire-management operations and the field reference data used to validate the Vegetation Fuel Type map. Given the scale and remoteness of many savanna projects, the site visit may be combined with the ground-truthing of that map.

Additional site visits may be required if there are substantial changes to project operations, or if deemed necessary by Isometric or the VVB. Site visit plans are determined by the VVB's internal assessment, in consultation with Isometric.

The VVB must comply with the requirements of the Isometric Standard and maintain and demonstrate expertise relevant to savanna fire management, including the remote sensing of fire and burned area, and savanna ecology and fuel assessment.

Savanna fire management is often a multi-party process, with burning, monitoring and benefit-sharing managed by different communities, operators or agencies.

Projects must demonstrate additionality in accordance with the Isometric Standard.

The counterfactual scenarios and baselines used to assess additionality are project-specific and are described in Section 8.0. For emission reductions, the counterfactual is the project area's own historical mean fire emissions, while removals use a dynamic matched-control baseline of woody-biomass change in a comparable control area.

Additionality is determined at Validation and reviewed at each Verification, and whenever project operating conditions change significantly, such as:

• a change in the legal obligations bearing on the activity; or
• project financials indicating that Carbon Finance is no longer required to sustain the managed fire regime.

If a review finds the Project to be non-additional, it becomes ineligible for future Credits; Certificates issued for current or past Crediting Periods are not affected.

Common practice analysis tests whether the project activity is already widespread without carbon finance.

Projects must conduct the analysis in accordance with the Isometric Standard, using the method for distributed technologies, with the following specifications.

In a savanna, fire itself is not the project activity and most eligible landscapes burn regularly, predominantly in the late dry season, both globally and in the launch regions^3,8. The project activity is a fire regime in which early dry season burning predominates.

Projects must account for uncertainty in the net CO₂e impact and determine the net impact for each Reporting Period conservatively, in accordance with the Isometric Standard. The quantity credited is the 30^th-percentile lower bound of the net estimate, after propagating the Project's measurement uncertainties. The error-propagation method is set out in Section 9.5.

Projects must provide a sensitivity analysis showing the effect of each input parameter's uncertainty on the final net estimate, in enough detail for a third party to reproduce it. An input variable may be omitted from the analysis where it contributes less than 1% of the net reduction or removal and all omitted variables together contribute less than 1% (see Section 8.0).

In accordance with the Isometric Standard, all evidence and data underlying the quantification of net reductions and removals, and the environmental and social safeguard monitoring, will be made available to the public through Isometric's platform. This includes:

• the Project Design Document;
• the GHG Statement;
• the annual burned-area maps;
• the total burned area by season and the proportion of the project area burned each year, alongside the Historical Reference Period mean;
• the Vegetation Fuel Type map;
• the years-since-last-burnt record and the fire-regime metrics used for the biodiversity safeguard;
• the woody-biomass measurements and the Donor Zone data used for the removals baseline;
• the emission factors and fuel parameters applied;
• the scientific literature used; and
• proof of approval for any necessary permits.

The Project Proponent may request that information subject to confidentiality be restricted to authorized Buyers, the Registry and the VVB; this does not apply to the numerical data produced or used in quantifying the net impact. In line with the FAIR principles (Findable, Accessible, Interoperable, Reusable)²³, the Project Proponent should also publicly disseminate deployment data relevant to scientific research.

The scope of this Protocol includes the GHG sources, sinks and reservoirs (SSRs) associated with a seasonal savanna fire management Project.

Projects must prepare a cradle-to-grave GHG Statement covering the GHG emissions of all activities within the system boundary.

GHG emissions associated with the Project may be direct emissions from a process, or indirect emissions from the combustion of fuels, electricity generation or other sources. The emissions of establishing, operating and decommissioning the Project are additional emissions in the project scenario and are deducted. Leakage, the displacement of fire beyond the Project Boundary, is monitored and deducted from the reductions account, as set out in Section 8.3. Projects must include all GHG SSRs within the system boundary, from the manufacture and transport of equipment and materials, through the operation of fire management, to the disposal of equipment at end of life, including the embodied emissions of equipment and consumables used in the Project. Projects must identify all sources of emissions directly or indirectly related to project activities.

The Project’s impact is quantified on two accounts within this single system boundary: a reductions account for the avoided CH₄ and N₂O emissions of savanna fire, and a removals account for the CO₂ stored in woody biomass. The two accounts are kept separate and non-overlapping, so that a fire year is charged correctly on each side, through different gases, without double counting.

The reductions account covers CH₄ and N₂O emissions of savanna fire, which occur in both the baseline and the project scenario. Only the incremental change attributable to the project activity is credited, as the baseline-minus-project difference (see Section 9.0). CO₂ from combustion is excluded, because it is biogenic and balanced by wet-season regrowth⁷. Carbon monoxide is excluded in this version, because no validated emission factor exists across the eligible regions (see Section 3.0).

The removals account covers the change in carbon stored in living woody biomass; soil organic carbon and pyrogenic carbon are excluded, because the evidence for a creditable, durable change is weak and conditional^24,25

Projects must include in the system boundary all relevant GHG SSRs controlled by and related to the Project, including but not limited to those in Table 1. Projects may exclude an SSR listed in Table 1 only with robust justification and evidence in the Project Design Document.

A SSR may be excluded from quantification as de minimis where a contribution analysis demonstrates that it accounts for less than 1% of the net CO₂e reduction or removal, and that all excluded SSRs together account for less than 1%, in accordance with the Isometric Standard and the GHG Accounting Module. This route applies in particular to the project establishment, operational and end-of-life emissions. The Materiality threshold for verification, across the totality of omissions, errors and misstatements, is 5% of the net reduction or removal (see Section 7.2.1).

Table 1. Scope of activities and GHG Sources, Sinks, Reservoirs (SSRs) to be included in the system boundary.

Projects must report any GHG SSR not captured by the categories in Table 1 as miscellaneous emissions. Projects must, in line with the GHG Accounting Module:

• consider all GHGs associated with their SSRs, including carbon dioxide, methane, nitrous oxide and, where present, fluorinated gases; for the CO₂ stored in woody biomass only CO₂ is quantified, while for all other activities all GHGs are considered;
• quantify emissions in tonnes of CO₂e using the 100-year global warming potential from the most recent IPCC Assessment Report (currently the Sixth); and
• consider the Materiality of SSRs in accordance with the Isometric Standard.

The baseline represents the emissions and the woody-biomass trajectory that would have occurred in the absence of the project activity. This Protocol uses a static baseline for the reductions account and a dynamic matched-control baseline for the removals account.

The following section outlines the standardized workflow that Isometric will take; the Project Proponent is not responsible for carrying out the steps in this section.

The reductions baseline is the project area's own mean annual fire emissions over the Historical Reference Period, computed before project activities begin and fixed for the Crediting Period (see Section 9.2.2). It is not adjusted for regional fire conditions during the Crediting Period, and it is not reassessed at renewal, because once the Project begins the project area no longer shows what would have happened without it.

Year-to-year variation in the historical record is carried into the uncertainty estimate (see Section 9.5) rather than into the baseline itself.

Fire weather is projected to become more severe across savanna regions as the climate warms, lengthening fire seasons and increasing the frequency of conditions that favour intense late dry season fire^26,27. A baseline set from historical emissions is therefore expected to understate the fire emissions that would occur without the Project over the Crediting Period, which makes it conservative.

In any Reporting Period in which the reductions account is negative, because the Project's measured emissions exceed the baseline or because attributed leakage exceeds the period's gross reductions, the shortfall is a reductions debit and is drawn from the Reductions Escrow (see Section 10.3.4). Where a reductions debit exceeds the balance of the Reductions Escrow, the remainder is carried forward and deducted from future Reduction Certificates before any further Reduction Certificates are issued.

The removals baseline is the counterfactual woody-biomass trajectory that would have occurred without the Project. It uses a pixel-matched process that compares the project's measured woody-biomass change against control pixels drawn from a Donor Zone (see Section 9.3.2.1). The credited removal is the project's change above this control. The control pixels are matched on the relevant environmental covariates, exclude the Project Boundary and the Leakage Monitoring Zone, and are reviewed at each verification for continued suitability. The detailed matched-control mechanics are set out in Section 9.3.2.

Savanna fire management is structurally exposed to leakage because it does not remove the cause of fire. The project changes how and when fire is applied inside the Project Boundary, but in southern African savannas the ignition agent is typically human and mobile^8,20, so suppressing fire inside the boundary can relocate it rather than prevent it. A pastoralist who can no longer burn inside the project area, or a hunter who used fire to flush game, may be incentivized to light fire outside the project area. This may be considered activity-shifting leakage, and a project that merely relocates fire, or shifts its timing without reducing total burned area, delivers marginal climate benefits³.

The project is not expected to take land out of production. Grazing, hunting and resource harvesting continue, and early dry season burning generally supports them, so there is no foreclosed output that simply relocates and market leakage is treated as zero, provided the project does not reduce fixed-site productivity. The residual risk is activity-shifting displacement, which the Protocol monitors.

Where this cannot be demonstrated, market leakage cannot be treated as zero, and no Certificates are issued for the Reporting Period until the reduction is remedied.

Projects must monitor a Leakage Monitoring Zone for potential activity-shifting leakage.

The Leakage Monitoring Zone is a band 5 km wide extending outward from the Project Boundary, including across any international boundary. The width is fixed rather than scaled to the project area, because the distance over which fire use is displaced depends on how far people travel to burn, not on the size of the project; the zone's area therefore grows with the length of the Project Boundary, where displacement occurs. It is restricted to eligible fuel and the same Vegetation Fuel Types as the project.

The Leakage Monitoring Zone is benchmarked against its own pre-project historical mean annual fire emissions and the standard deviation of those emissions, and is monitored each Reporting Period. Emissions are quantified across both seasons, using the same emissions stack, seasonal classification rule and Vegetation Fuel Types as the project area. Historical emissions in the Leakage Monitoring Zone are calculated exactly as the project area's historical emissions are, and Reporting Period emissions exactly as the project area's Reporting Period emissions are, so that a fire burning across the Project Boundary is treated the same on both sides of it.

Fire emissions vary widely from year to year, and displaced fire cannot be separated cleanly from that natural variation. Only emissions above the Leakage Monitoring Zone's historical mean plus one standard deviation are therefore attributed to the Project as leakage: years within that range are treated as consistent with the zone's historical fire regime, and only the excess above the threshold is deducted (see Section 9.4.4). Any reduction of fire in the zone that results from the Project lowers the chance of exceeding the threshold, but is never credited.

Attributed leakage is deducted from the reductions account of the Reporting Period in which it occurs, and never from removals (see Section 9.1 and Section 9.4.4). Where it exceeds that period's gross reductions, the shortfall is drawn from the Reductions Escrow and then carried forward, as set out in Section 10.3.4. A reduction in fire in the Leakage Monitoring Zone might reflect good management spilling outward and would be considered positive leakage, but is not credited consistent with conservative accounting.

The Leakage Monitoring Zone is exclusive from the project area and the baseline donor zone without overlap. Excluding the Leakage Monitoring Zone from the donor is essential, because displaced project fire inside the Donor Zone would corrupt the very baseline. Community-led fire management (see Section 6.0) is the upstream control on leakage, because integrating the people who set fire as the project's fire managers removes the motivation for ignition⁹.

This section sets out how net emission reductions and net removals are calculated for each Reporting Period. The two are quantified on separate ledgers and issued as separate Certificates. A Reduction Certificate represents one tonne of CO₂e of reduced methane and nitrous oxide emissions. A Removal Certificate represents one tonne of CO₂e stored in woody biomass. All quantities are expressed in tonnes of CO₂e.

Each Reporting Period, the Project computes gross emission reductions, gross removals, the project's life-cycle emissions, and leakage, and from these the net reduction and net removal that are issued as Credits.

Reductions and removals each net against their own counterfactual. The reductions counterfactual is the historical mean fire emissions baseline (see Section 8.2.1), and the removals counterfactual is a matched control of changing biomass within the region (see Section 8.2.2). The removals account is calculated using a cumulative framework, which evaluates removals from project initiation to the end of the current Reporting Period, and then issues new Removal Certificates for the current Reporting Period corresponding to the increase in net removals relative to what was previously awarded.

The project's life-cycle emissions form the project emissions term (Equation 12). They are mostly fossil CO₂, and are deducted from the removals account first, reducing the reductions account only to the extent they exceed the Reporting Period's gross removals. In a Reporting Period in which removals are not quantified, gross removals are zero for this purpose, and that period's life-cycle emissions are deducted from the reductions account.

Leakage is measured as methane and nitrous oxide from displaced fire, so it is deducted from the reductions account alone and never from removals (see Equation 1 and Section 9.4.4). A negative reductions account in a Reporting Period is a reductions debit, drawn from the Reductions Escrow as set out in Section 10.3.4.

The quantity issued for each account is the conservative lower bound of the net result, as set out in Section 9.5. Projects must select the value that conservatively limits the credited result where data are thin or a parameter is ambiguous.

The reductions account Certificates the avoided methane and nitrous oxide emissions of savanna fire, issued separately from the Removal Credits. Gross emission reductions for the Reporting Period are the difference between the baseline fire emissions and the project's measured fire emissions, less the attributed leakage:

$$CO_2e_{Reductions,RP} = CO_2e_{Fire,Baseline,RP} - CO_2e_{Fire,Project,RP} - CO_2e_{Leakage,RP}$$

(Equation 1)

Where:

• $CO_2e_{Reductions,RP}$ represents the emission reductions net of attributed leakage, as CH₄ and N₂O, for the Reporting Period, $RP$, in tonnes of CO₂e.
• $CO_2e_{Fire,Baseline,RP}$ represents the baseline fire emissions for the Reporting Period, $RP$, in tonnes of CO₂e.
• $CO_2e_{Fire,Project,RP}$ represents the project's measured fire emissions for the Reporting Period, $RP$, in tonnes of CO₂e.
• $CO_2e_{Leakage,RP}$ represents the attributed leakage for the Reporting Period, $RP$, in tonnes of CO₂e (see Section 9.4.4).

The Project's life-cycle emissions are deducted from the removals account first, and reduce the reductions account only to the extent they exceed the Reporting Period's gross removals. Net emission reductions are therefore the gross reductions, less only the part of the project emissions that the period's removals could not absorb.

A negative $CO_2e_{Reductions,RP}$ is a reductions debit, drawn from the Reductions Escrow (see Section 10.3.4).

The project's fire emissions for the Reporting Period, $CO_2e_{Fire,Project,RP}$, and the fire emissions of each historical year, $CO_2e_{Fire,y}$ (see Section 9.2.2), are computed with the same equation, applied respectively to the recent record and to the historical pre-project record. Fire emissions are the sum, across Vegetation Fuel Types, seasons and fuel classes, of the burned area, the fraction of it that actually burnt, the fuel load, the combustion completeness, and the emission factor and global warming potential of each gas:

$$CO_2e_{Fire}=\sum\limits_s\sum\limits_v\sum\limits_f(A_{v,s} \times P_{v,s} \times FL_{v,f} \times CC_{s,f} \times M_s \times \sum\limits_g(EF_{g,f} \times GWP_g))$$

(Equation 2)

Where:

• $s$ represents the season (early or late dry season, see Section 4.2.4)
• $v$ represents each Vegetation Fuel Type (see Section 12.1)
• $f$ represents the fuel class
• $A_{v,s}$ represents the mapped burnt area in Vegetation Fuel Type $v$ and season $s$, in hectares
• $P_{v,s}$ represents the patchiness factor for Vegetation Fuel Type $v$ and season $s$. This is the dimensionless fraction of the mapped scar that actually burnt. Its default values are given for the project's region in Section 12.4 (Table A6).
• $FL_{v,f}$ represents the fuel load of fuel class $f$ (fine, shrub, coarse, heavy) in Vegetation Fuel Type $v$, in tonnes of dry matter per hectare. Its default values are given for the project's region in Section 12.2 (Tables A2 to A4).
• $CC_{s,f}$ represents the dimensionless combustion completeness for season $s$ and fuel class $f$. Its default values are given for the project's region in Section 12.3 (Table A5).
• $EF_{g,f}$ represents the emission factor for gas $g$ (CH₄, N₂O) from fuel class $f$, in tonnes of gas per tonne of dry matter combusted. Its default values are given for the project's region in Section 12.5 (Tables A7 and A8).
• $M_s$ represents the uncured-burn multiplier, equal to 2 for uncured burns in a Reporting Period (see Section 4.2.4), which take the late dry season parameters, and 1 otherwise.
• $GWP_g$ represents the 100-year global warming potential for gas $g$, from IPCC AR6²⁸: 27 for CH₄ and 273 for N₂O.

Projects must convert any emission factor reported as a fraction of fuel carbon or nitrogen, as the regional parameter sets in Section 12.5 are, to a mass of gas per unit of dry matter using the fuel carbon and nitrogen contents given there and the molecular mass ratios of methane to carbon and nitrous oxide to nitrogen.

The effective area burnt in each stratum is the mapped fire-scar area multiplied by the patchiness factor. The patchiness factor is the fraction of the mapped scar that actually burnt. It is applied as a per-Vegetation-Fuel-Type, per-season default from Section 12.4 rather than measured for each fire. The lower burnt fraction of early dry season fire is one of the main levers of abatement, alongside reduced combustion completeness.

Projects must map every burn in a year, not only the first. On the project side, each burn is counted as a separate fire and classified by its own date, so ground burnt twice in a year contributes the emissions of both fires. In the historical record, only the first burn of each pixel in a year is counted.

Fuel load is determined per Vegetation Fuel Type and fuel class. For fine and shrub fuels, which accumulate measurably with time since fire, the fuel load is a function of years since last burnt:

$$ln(FL(f)) = a(f)+b(f)\times ln(YSLB)$$

(Equation 3)

Where:

• $FL(f)$ represents the fuel load of fuel class $f$ (fine, coarse, heavy or shrub), in tonnes of dry matter per hectare. Grass and litter typically dominates the fine-fuel load and varies between early and late dry seasons.
• $YSLB$ represents the years since the stratum last burnt. Years since last burnt is never taken as less than one, so ground burnt again in the same year is assigned the one-year fuel load for the season of the later burn.
• $a(f)$ and $b(f)$ represent the coefficients of the time-since-fire relationship for fine and shrub fuels. Section 12.2 tabulates, for the project's region, the fuel-load values this relationship yields by years since last burnt.

Projects must update the years-since-last-burnt map each Reporting Period from a running burned-area record of at least five years, setting burnt pixels to age zero and incrementing the rest.

The Project's coarse and heavy woody-fuel load is taken from the most recent measurement of the woody-biomass stock used in the removals account, so one physical pool feeds both accounts. When coarse and heavy woody-fuel loads are not available, the Project may use the pooled mean loads given in Section 12.2.

Combustion completeness is the fraction of available fuel class $f$ consumed by fire. It is applied as a fixed season-by-class default per Vegetation Fuel Type $v$, given for the project's region in Section 12.3, on both the baseline and the project side. The lower completeness of early dry season fire is a core part of the per-fire emission reduction.

Emission factors are field-measured Tier 2 values per Vegetation Fuel Type $v$, given for the project's region in Section 12.5, each reported with its sample size and error. Direct measurements find no significant seasonal difference in the methane or nitrous oxide emission factors in northern Australia, and mostly none in miombo^3,4,29; the seasonal contrast is carried by combustion completeness and patchiness, not by the emission factor.

This season-invariance applies to cured fuels. Fuels burnt before they have cured burn with lower combustion efficiency and a higher methane yield^17,18,19, so such burns are treated as uncured burns under Section 4.2.4.

Early dry season fires consume little live canopy or woody fuel. In miombo, early dry season combustion completeness is 0.05 to 0.06 for shrub fuels and 0.01 for heavy fuels⁴, so nitrogen-rich green foliage contributes little to nitrous oxide emissions from early dry season fire.

The Project's counterfactual baseline fire emissions, $CO_2e_{Fire,Baseline}$, are the project area's mean annual fire emissions over the Historical Reference Period:

$$CO_2e_{Fire,Baseline}=\frac{1}{n}\sum\limits_{y=1}^{n} CO_2e_{Fire,y}$$

(Equation 4)

Where:

• $CO_2e_{Fire,y}$ represents the project area's fire emissions in historical year $y$, in tonnes of CO₂e, calculated with Equation 2 (see Section 9.2.1).
• $n$ represents the number of years in the Historical Reference Period.

This value is calculated once before project activities begin and fixed for the Crediting Period.

The removals account credits the increase in carbon stored in woody biomass above a matched control, issued separately from the Reduction Credits. The carbon pools are living aboveground and belowground woody biomass, which can be quantified accurately and monitored over time; soil organic carbon, litter and pyrogenic carbon are excluded (see Section 8.0).

Net removals for the current Reporting Period, $RP_n$, are calculated as:

$$CO_2e_{Removals,RP_n} = CO_2e_{Biomass,Stored,cum} - CO_2e_{Biomass,Counterfactual,cum} - CO_2e_{Emissions,RP_n} - \sum_{i=1}^{n-1} CO_2e_{Removals,RP_i}$$

(Equation 5)

Where:

• $CO_2e_{Removals,RP_n}$ is the net carbon removals in woody biomass for the current Reporting Period, $RP_n$, in tonnes of CO₂e.
• $CO_2e_{Biomass,Stored,cum}$ is the total cumulative CO₂ removed from the atmosphere and stored as organic carbon in living woody biomass since project initiation, in tonnes of CO₂e (see Section 9.3.1).
• $CO_2e_{Biomass,Counterfactual,cum}$ is the total cumulative counterfactual CO₂ removed from the atmosphere and stored as organic carbon in living woody biomass in the absence of project activities since project initiation, in tonnes of CO₂e (see Section 9.3.2).
• $CO_2e_{Emissions,RP_n}$ is the project's life-cycle emissions for the current Reporting Period, $RP_n$, in tonnes of CO₂e (see Section 9.4).
• $\sum_{i=1}^{n-1} CO_2e_{Removals,RP_i}$ is the sum of net removals already credited to the Project for all prior Reporting Periods, in tonnes of CO₂e. These values are those determined at the time of each prior Verification and are treated as fixed; they are not recalculated using currently available data.

The cumulative woody-biomass carbon stock change at time $t$ is:

$$CO_2e_{Biomass,Stored,cum}=\left(M_{Biomass}(t)-M_{Biomass}(0)\right) \times CF \times \frac{44}{12}$$

(Equation 6)

Where:

• $M_{Biomass}(t)$ and $M_{Biomass}(0)$ are the total woody biomass at time $t$ and at project initiation, respectively, in tonnes of dry biomass.
• $CF$ is the average fraction of carbon content for the tree species in the project area, in $\frac {tonne: C}{tonne: dry: biomass}$.
• $\frac{44}{12}$ is the ratio of the mass of CO₂ to the mass of C, used to convert carbon to CO₂e.

Projects must choose the carbon fraction, $CF$, from the following hierarchy:

1. A regional and species-specific factor that is justified based on scientific literature (e.g., Doraisami et al., 2022³⁰). This is the preferred approach to have the most accurate estimate and avoid overestimation;
2. If the above is not available, then a genus-specific or national average factor that is justified based on scientific literature can be used;
3. As a last resort, if it is proven that the above two factors are not available, then a default factor of 47%, which is a mean across species^31,32, can be used.

The cumulative carbon stored in woody biomass is calculated by taking the difference between project initiation and time $t$. At each time point, the biomass is calculated as:

$$M_{Biomass}=M_{AGB}+M_{BGB}$$

(Equation 7)

Where:

• $M_{AGB}$ represents the total aboveground woody biomass, in tonnes of dry biomass.
• $M_{BGB}$ represents the total belowground woody biomass, in tonnes of dry biomass.

The carbon stored in belowground woody biomass is derived from the aboveground biomass by a root-to-shoot ratio:

$$M_{BGB}=M_{AGB}\times RS$$

(Equation 8)

Where:

• $RS$ is the root-to-shoot ratio, which is a dimensionless belowground biomass to aboveground biomass ratio.

Appropriate root-to-shoot ratios should be selected by regional and species-specific factors that are justified based on scientific literature. This is the preferred approach to have the most accurate estimate and avoid overestimation.

Where no suitable project-specific factor can be obtained, Projects must use the root-to-shoot ratio for the ecological zone and continent of the project area from IPCC 2019 Table 4.4³¹, and provide evidence of the unsuccessful search for project-specific factors, showing:

• a list of search terms used within a research database (e.g., Web of Science, Google Scholar) that encapsulate the region and species relevant to the Project; and
• the relevant species are not included in the list of species for which root-to-shoot ratio data are available on the TRY Plant Trait Database³³.

Projects must report the uncertainty of the selected root-to-shoot ratio from the same source dataset. For example, the IPCC 2019 Chapter on Forest Land³¹ provides an uncertainty in the root-to-shoot ratio.

This Protocol currently supports the following three Capture and Conversion Modules for quantifying the total AGB over the project area at a point in time, $M_{AGB}$:

Requirements for each approach are described in the corresponding Modules. Projects must state in the PDD which Module is used and meet its requirements. Note that Projects using LiDAR and Earth Observation Maps for quantification still require field plots as the source of truth for benchmarking the maps. This list of acceptable approaches may be expanded upon in future versions of the Protocol.

This Protocol uses a dynamic baseline approach to quantify the counterfactual impact on woody carbon stocks if the project activity had not occurred for quantifying Removal Certificates stored in biomass. In this approach, the counterfactual is determined by observing changes in woody carbon stocks for a collection of areas outside of the project area (control pixels) that are representative of the project area except for the project activity. These control pixels are not used for determining the counterfactual for estimating fire emission reductions.

By using observations of matched controls, dynamic baselines are able to reflect changes in market trends, policies, environmental changes, etc., that can affect counterfactual carbon storage and which would be difficult to capture in static approaches. As such, the use of real-time remote sensing and robust matching procedures in the dynamic baseline procedure leads to the most plausible baseline scenario that can be clearly quantified and compared to the project activities. Further, the pixel matching procedure matches every pixel within the project area to multiple pixels in the control area, generating an ensemble of samples representing multiple baseline scenarios. This ensemble approach inherently characterizes uncertainty probabilistically through the variation in control pixels. This uncertainty is then included in carbon calculations (see Section 9.5). Because of this, the use of dynamic baseline approaches that leverage remote sensing to compare project activities to matched controls has been noted as a rigorous and conservative approach in the scientific literature^34,35,36,37.

Dynamic baselines will be independently determined and transparently reported by Isometric at each Verification to determine any deduction in Certificate issuance based on the Baseline scenario. Certificate issuance will only occur for carbon removal that is determined to be additional via the following procedure, inclusive of uncertainty.

The following section outlines the standardized workflow that Isometric will take. The Project Proponent is not responsible for carrying out the steps in this section. Project Proponents may suggest areas that could constitute suitable control pixels or features for matching based on their expert knowledge of their unique system. However, the ultimate determination of control pixels will be done by Isometric following the procedure and criteria below. Although dynamic baseline approaches are reliant on the suitability of the matched areas to act as controls, the standardized approach includes provisions for using several criteria for matching, matching to multiple pixels, assessing match quality, expanding the number of potential matches, and regularly reassessing control pixel suitability to minimize the associated uncertainty.

Isometric will select control pixels from a Donor Zone that meets the following eligibility criteria:

• located in the same ecoregion and Vegetation Fuel Type as the project area;
• excludes the Leakage Monitoring Zone (see Section 8.3);
• does not lie within other carbon projects, or within areas under a documented early dry season fire-management programme, whether or not it receives carbon finance;
• does not contain unrepresentative land use types (e.g., urban areas);
• is subject to the same or similar relevant regulations, government incentives, and programs as the project area; and
• shows no statistically significant difference (p ≥ 0.05) from the project area across the carbon proxy.

Within The Donor Zone, project and control pixels are matched on the environmental covariates that govern woody-biomass accumulation and fire response:

• Vegetation Fuel Type
• Pre-project biomass trajectory
• Bioclimatic variables (e.g., temperature and precipitation)
• Topography

Initially, the potential area for the Donor Zone should be limited to a 50 km band beyond the Leakage Monitoring Zone. However, if suitable matches (see Section 9.3.2.3 for matching step) are not found in this zone, additional step-outs in 10 km increments may occur to find appropriate control pixels, assuming they meet the criteria above.

Once the boundaries of the Donor Zone are determined, Isometric will generate high-resolution (≤30 m) pixel maps representing woody carbon stocks or a suitable proxy for carbon stocks. At project initiation, the layers will cover the entire project area and Donor Zone at the same resolution for at least five historical time points, each separated by at least 1 year.

Isometric will select a suitable proxy that meets the following criteria:

• The proxy is correlated with AGB throughout the Crediting Period.
• If the proxy saturates in the project area or Donor Zone during the Crediting Period and becomes insensitive to further carbon gain (e.g., as is the case for many simple optical indices), the Project will be transitioned to a new proxy better suited for the current project conditions (see Section 9.3.2.5)³⁸. Proxies such as canopy height from models which use datasets from multiple types of earth observation (particularly 2D satellite imagery, LiDAR and SAR) are preferred^39,40.
• The proxy map product should exclude cloud occlusion, saturation, and other contamination. This may lead to seasonal composite images.
• The proxy should be insensitive to seasonal variability, which may be addressed through taking an annual average to remove the seasonal cycle.
• The proxy should be stable, so that in the absence of carbon change, the proxy value stays the same.
• If an AGB map is used instead of a proxy map, the AGB map product used to generate the dynamic baseline will meet the quality requirements in the corresponding Module. If an AGB map is also being used for AGB quantification, the same data product should be used for both the baseline and quantification, when possible (see Section 9.3.1.1).

Project pixels are matched to control plot pixels based on the historical time series of the selected carbon proxy, $C_{proxy}$, of each pixel, using k-nearest neighbors with replacement (or an alternative justified algorithm). This matching will use, at a minimum, five historical time points capturing at least the five years prior to project initiation. Each project pixel will be matched to a minimum of 10 different control pixels, and the mean carbon proxy over the group of control pixels is calculated from the map product created using the procedure described in Section 9.3.2.2. Multiple project pixels may be matched to the same control pixels.

For each project pixel, the change in carbon stock since project initiation is calculated both for the project pixel and for the collection of corresponding control pixels (taking the mean across the group) using the values from the carbon proxy map:

$$\Delta C_{proxy,t}=C_{proxy}(t)-C_{proxy}(0)$$

(Equation 9)

Where:

• $C_{proxy}(t)$ and $C_{proxy}(0)$ are the proxy values at times $t$ and project initiation, respectively.
• $\Delta C_{proxy,t}$ is the change in carbon proxy value from project initiation to time $t$.

The carbon removal of the counterfactual scenario is found by scaling the quantified carbon removal in the project area by the ratio of the mean proxy change in the matched control pixels and the mean proxy change in the project pixels:

$$CO_2e_{Biomass,Counterfactual,cum}=CO_2e_{Biomass,Stored,cum} \times \frac{\frac{1}{N}\sum_{i=1}^{N}\max\left(0,\frac{1}{k_i}\sum_{j=1}^{k_i}\Delta C_{proxy,control,ij}(t)\right)}{\frac{1}{N}\sum_{i=1}^{N}\Delta C_{proxy,project,i}(t)}$$

(Equation 10)

Where:

• $CO_2e_{Biomass,Counterfactual,cum}$ is the total counterfactual CO₂ removed from the atmosphere and stored as organic carbon in living woody biomass in the absence of project activities since project initiation, in tonnes of CO₂e.
• $CO_2e_{Biomass,Stored,cum}$ is the cumulative CO₂e stored in living woody biomass since project initiation, as calculated in Section 9.3.1.
• $\Delta C_{proxy,project,i}(t)$ is the cumulative proxy change at time $t$ in project pixel $i$, of $N$ total.
• $\Delta C_{proxy,control,ij}(t)$ is the cumulative proxy change at time $t$ in control pixel $j$, of $k_i$ total, matched to project pixel $i$.
• $\Delta C_{proxy,control,i}(t)$ is the mean of $\Delta C_{proxy,control,ij}(t)$ across the $k_i$ control pixels matched to project pixel $i$.

In Equation 10, the mean proxy change in the ensemble of control pixels matched to a given project pixel is floored at 0. This is done to ensure that only removals are credited under this Protocol.

Isometric will test the additionality condition, that per-pixel differences $\Delta C_{proxy,project,i}(t)-\Delta C_{proxy,control,i}(t)$ are statistically greater than zero across all project pixels, using a paired, one-sided t test (p < 0.05, inclusive of uncertainty). If the condition is not met, no Removal Certificates are issued at time $t$, but previously issued Removal Certificates are not affected.

The counterfactual carbon storage is then used to calculate a performance benchmark for the project area, $PB_{cum}$:

$$PB_{cum}=\frac {CO_2e_{Biomass,Stored,cum}}{CO_2e_{Biomass,Counterfactual,cum}}$$

(Equation 11)

If the Project is additional, the performance benchmark will be greater than 1, with larger magnitudes indicating a greater difference between storage in the Project and control areas.

At each Verification, the control pixels are reviewed to determine continued eligibility against the criteria outlined in Section 9.3.2.1. Changes in the status of these criteria which create a divergence between the control pixel and the project pixel indicate that the pixel may no longer be a suitable control for the project area. As an example, if a control area comes under a documented early dry season fire-management programme, but the project area's circumstances do not change in the same way, the control pixel would no longer be considered representative. In the event that control pixel matches are no longer suitable, replacements will be selected for the impacted project pixels. Control pixels may also be replaced if they experience persistent data quality issues. Any replacement of control pixels will be clearly documented and justified by Isometric, and the replacement pixels are then used for the full cumulative counterfactual calculation.

Over the course of the Crediting Period, there may also be updates to the proxy dataset used for the dynamic baseline assessment. Isometric will make such changes independently of the resulting counterfactual calculation, and only where there is a justified need for changing the carbon proxy, which may include:

• A discontinuation of the current product
• Saturation of the current proxy with woody-biomass growth
• A demonstrated improvement in the proxy's ability to capture aboveground biomass in the project area (e.g., new data products become available)

Isometric will document and justify any proxy replacement, and will only adopt a new proxy with equivalent or better correlation with field-measured aboveground biomass in the project area. The application of the new carbon proxy product will follow the below hierarchy:

1. Where the new product coverage includes the five-year pre-project window forward, the new proxy is used to re-match the control pixels and for calculation of the cumulative counterfactual over the full Crediting Period.
2. Where the new product coverage includes project initiation, the controls are not re-matched but the proxy is used for the calculation of the cumulative counterfactual over the full Crediting Period.
3. Where the new product coverage does not include the full Crediting Period, it may only be adopted if it has at least two time points of overlapping coverage with the previous proxy. The two products are harmonized over the overlapping data period to enable a conversion of the prior product's values over the initial period where new proxy data are not available. This harmonized proxy is then used for the calculation of the cumulative counterfactual over the full Crediting Period. Control pixels are not re-matched in this scenario.

Across all of the above scenarios, previously issued Removal Certificates are not impacted by the change in proxy (i.e., the final summation term in Equation 5).

The Project emissions term is the project's life-cycle greenhouse gas emissions:

$$CO_2e_{Emissions,RP}=CO_2e_{Establishment,RP}+CO_2e_{Operations,RP}+CO_2e_{End-of-Life,RP}$$

(Equation 12)

Where:

• $CO_2e_{Establishment,RP}$ represents the emissions of project establishment allocated to the Reporting Period, $RP$, in tonnes of CO₂e.
• $CO_2e_{Operations,RP}$ represents the emissions of fire-management operations in the Reporting Period, $RP$, in tonnes of CO₂e.
• $CO_2e_{End-of-Life,RP}$ represents the end-of-life emissions allocated to the Reporting Period, $RP$, in tonnes of CO₂e.

These emissions are included where material, subject to the de minimis threshold set in the Section 8.0.

Establishment emissions, $CO_2e_{Establishment,RP}$, arise from project inception to the moment before the first fire-management activity, including but not limited to the SSRs set out in Table 1, such as the embodied, transport and installation emissions of equipment, materials and firebreak works. GHG emissions associated with project establishment may be amortized over the anticipated project lifetime, or per output of product. Rules on amortization are outlined in Section 7 of the GHG Accounting Module.

GHG emissions associated with $CO_2e_{Operations,RP}$ should include all emissions associated with operational activities, including but not limited to the SSRs set out in Table 1. Operations emissions are the fossil emissions of fire management in the Reporting Period, such as fuel used in aerial ignition and in vehicles. Projects must attribute operations emissions to the Reporting Period in which they occur. Allocation outside of the current Reporting Period may be permitted in certain instances, on a case by case basis in agreement with Isometric.

End-of-life emissions, $CO_2e_{End-of-Life,RP}$, includes all emissions associated with activities that are anticipated to occur after the Crediting Period, including the disposal of equipment and activities related to ongoing monitoring for Reversals of the removals stock.

Projects must estimate $CO_2e_{End-of-Life,RP}$ upfront and allocate it in the same way as $CO_2e_{Establishment,RP}$.

Projects must revisit the end-of-life assumptions at each Reporting Period and make any necessary adjustments.

Leakage is the project-attributable increase in fire emissions in the Leakage Monitoring Zone, above what the area would have emitted without the project. Fire emissions in the Leakage Monitoring Zone are quantified with Equation 2, across both seasons and all Vegetation Fuel Types and fuel classes, using the same burned-area product, seasonal classification rule and parameter set as the project area. Historical emissions in the Leakage Monitoring Zone are calculated in the same way as the project area's historical emissions (see Section 9.2.2), and Reporting Period emissions in the same way as the project area's Reporting Period emissions (see Section 9.2.1). Leakage is the part of a Reporting Period's emissions in the zone that exceeds its historical mean by more than one standard deviation of its historical annual emissions:

$$CO_2e_{Leakage,RP}=\max\left(0,\ CO_2e_{LMZ,RP}-\left(CO_2e_{LMZ,Historical}+\sigma_{LMZ,Historical}\right)\right)$$

(Equation 13)

Where:

• $CO_2e_{LMZ,RP}$ represents the measured fire emissions in the Leakage Monitoring Zone during the Reporting Period, $RP$, in tonnes of CO₂e.
• $CO_2e_{LMZ,Historical}$ represents the Leakage Monitoring Zone’s pre-project historical mean annual fire emissions, in tonnes of CO₂e.
• $\sigma_{LMZ,Historical}$ represents the sample standard deviation of the Leakage Monitoring Zone's annual fire emissions over the Historical Reference Period, in tonnes of CO₂e.

The historical mean is computed over the same window used for the Project's own historical baseline, being twice the mean fire-return interval or ten years, whichever is longer, and is fixed at project initiation. The standard deviation is computed over the same years and fixed at the same time.

Emissions in the Leakage Monitoring Zone at or below this threshold are treated as consistent with the zone's historical fire regime, because displaced fire cannot be separated cleanly from natural year-to-year variation. Only the excess above the threshold is deducted, from the reductions account of the Reporting Period (see Section 9.1).

A reduction in fire in the Leakage Monitoring Zone below its historical mean is positive leakage. It is documented but set to zero, never credited.

Projects must account for GHG emissions following the GHG Accounting Module, which sets the data-quality hierarchy, the materiality and de minimis tests, the emissions-amortization and allocation rules, and the treatment of embodied and transport emissions.

Projects must calculate energy-related emissions, such as fuel for aerial ignition and vehicles and electricity for monitoring equipment, following the Energy Use Accounting Module.

The quantity issued for each account is a conservative lower bound of the net result, not its central estimate. The credited quantity is the 30th-percentile lower bound of the net estimate, computed separately for reductions and for removals. The lower bound is obtained by propagating the Project's measurement uncertainties through the calculation, including:

• burned-area mapping error, from the documented accuracy of the method or product used;
• fuel load, including the woody-biomass measurement uncertainty that enters both the reductions and the removals accounts through the shared woody pool;
• combustion completeness and emission factors, with their reported errors;
• the carbon fraction and root-to-shoot ratio applied in the removals account; and
• variance and uncertainty in the counterfactuals, including the interannual variability of the historical reductions baseline and the matched-control variance of the removals baseline.

The storage reservoir for the CO₂ credited as a removal under this Protocol is living aboveground and belowground woody biomass, the same carbon pools quantified in Section 9.3. The durability of a removal is the length of time for which that CO₂ is kept out of the atmosphere and so cannot contribute to further warming. This section sets out the durability of the Removal Credits, the risks of Reversal, and the monitoring and compensation that protect them.

The durability and Reversal provisions of this section apply to the removals account only. An avoided methane or nitrous oxide emission, once avoided, cannot be re-emitted, so the reductions account carries no permanence obligation. Reductions debits are instead covered by a project-specific Reductions Escrow, which is not a Buffer Pool, as set out in Section 10.3.

The durability of a Removal Certificate is the length of the Ongoing Monitoring Period committed at PDD submission (see Section 5.4). No Removal Certificates are issued during the Ongoing Monitoring Period.

Projects must maintain the credited woody-biomass stock throughout the Project Commitment Period; Isometric monitors it for loss.

A savanna-fire-management-wide removals Buffer Pool managed by Isometric insures the Removal Certificates against Reversals. Throughout the Ongoing Monitoring Period, Isometric monitors for Reversals to ensure the Certificates achieve their stated durability, and on detection and quantification of a loss, Removal Certificates issued to the removals Buffer Pool are cancelled in equal proportion to the loss for which the Project Proponent is liable at the time of the Reversal (see Section 10.1.1, Section 10.3, and Section 10.4).

Project Proponents are liable for issued Removal Certificates for the duration of their durability, starting from the time of issuance. This rolling window of liability sets the time period for which a Reversal of stored carbon is subject to compensation via the removals Buffer Pool. For example, for a Project with a 40-year durability, the Project Proponent would be liable for potential Reversals in Removal Certificates issued in Year 1 of the Project for Years 2 through 41 of the Project. At any given point in the Project Commitment Period, the liable volume of Removal Certificates is the sum of all Removal Certificates for which the durability window is active. This liable volume represents the upper limit for compensation via the removals Buffer Pool. Removal Certificates for which the durability window has expired are considered to have already met their stated durability commitment and are no longer subject to treatment as a Reversal. When a Reversal occurs, the accounting conservatively follows a "first in, first out" approach, wherein the Reversal is applied first to the oldest Removal Certificates within the liable volume.

A Reversal is a measured loss of previously credited woody-biomass stock that returns CO₂ to the atmosphere. Unlike the avoided fire-emission flux, the accumulating woody stock is exposed to release by severe fire, drought or other disturbance⁵. For a savanna fire project, the dominant reversal pathway is a return of intense late dry season fire regime, whether through a lapse in management or repeated severe fire years, with drought as another mechanism.

The likelihood and severity of a Reversal are shaped by factors outside and inside the Project's control. External factors include regional climate change and drought trends, severe fire seasons and lightning ignition, encroachment or land-use change in and around the project area, rising ignition pressure in the surrounding landscape, and changes in herbivore pressure, including elephant impacts on woody biomass⁸. Project-related factors include the design and continuity of the managed fire regime, the strength of community fire governance and the integration of traditional fire users (see Section 6.0), the maintenance of firebreaks and early-season burning, and the Project's financial and institutional stability.

For a savanna fire project, the core reversal-mitigation activity is the managed fire regime itself. This is sustained early dry season burning that fragments the landscape and prevents severe late dry season fire from carrying through the woody stock.

A removals Buffer Pool, managed by Isometric and governed by the Isometric Standard, insures the Project's credited removals against Reversals. On detection and quantification of a Reversal, an equal quantity of Certificates is cancelled from the removals Buffer Pool, so that the net impact already credited remains whole. Reduction Certificates carry no permanence obligation and are not insured by a Buffer Pool. Reductions debits are not Reversals: they are covered by the project-specific Reductions Escrow (see Section 10.3.4) and are never compensated from the removals Buffer Pool.

Each Reporting Period, the Project contributes a flat 20% of the Removal Certificates generated that period to the removals Buffer Pool. The contribution is a fixed proportion and is not risk-rated; the rate does not vary with the reversal-risk rating required under Section 10.2 or with any change in the Project's risk profile over the Project Commitment Period.

The removals contributions are held in a savanna-fire-management-wide Buffer Pool managed by Isometric. The composition of the pool is transparently reported on the Isometric Registry.

The compensation process is governed by the Isometric Standard. On detection and quantification of a loss, the following apply.

• A Reversal of removals within the removals Crediting Period is incorporated into the removals quantification at the next Verification at which removals are quantified. If the net removal term (Equation 5) in a Reporting Period is found to be negative (woody-biomass carbon stock at $t$ < woody-biomass carbon stock at $t-1$), removals Buffer Pool Certificates are cancelled equal to the net emissions from the Reporting Period or the liable volume of Removal Credits, whichever is smaller. A Reversal of removals caused by a prescribed burn is an Avoidable Reversal.

• A Reversal of removals during the Ongoing Monitoring Period is quantified as described in Section 10.4 and compensated from the removals Buffer Pool within one year of the loss event. Compensation will equal the smaller of the quantified Reversal or the liable volume of Removal Certificates at the time of the Reversal.

• For an Avoidable Reversal, Isometric cancels Certificates from the removals Buffer Pool equal to the lesser of the quantified Reversal or the liable volume of Removal Certificates at the time of the Reversal. Projects must replenish them before further Removal Certificates are issued. Replenishment comes from Removal Certificates generated in the next Reporting Period in which removals are quantified during the Crediting Period. Replenishment comes from Certificates of the same type generated in the next Reporting Period during the Crediting Period. After the Crediting Period, replenishment comes from Certificates generated by other projects operated by the Project Proponent, or from Certificates of equivalent quality sourced at the Project Proponent’s expense and accepted by Isometric.

• For an Unavoidable Reversal, Isometric cancels Certificates from the removals Buffer Pool equal to the lesser of the quantified Reversal or the liable volume of Removal Certificates at the time of the Reversal.

Each Reporting Period, 5% of the Reduction Certificates generated that period are held in a project-specific Reductions Escrow on the Isometric Registry. Certificates in the Reductions Escrow cannot be Transferred or used to make Deliveries until they are released. The Reductions Escrow is not a Buffer Pool: avoided methane and nitrous oxide emissions cannot be re-emitted, and the escrow exists to keep the reductions account unbiased across the Crediting Period, because the fixed historical baseline over- and under-states each year's counterfactual emissions (see Section 8.2.1). The contribution is a flat 5% and is not risk-rated.

• A reductions debit arises in any Reporting Period in which the reductions account is negative, whether because the project area's fire emissions exceed the baseline or because attributed leakage exceeds the period's gross reductions (see Section 9.1). The debit is drawn from the Reductions Escrow, whatever its cause.
• Where a reductions debit exceeds the balance of the Reductions Escrow, the remainder is carried forward as a debt and deducted from the Project's future Reduction Certificates before any further Reduction Certificates are issued.
• At the end of each Crediting Period, the remaining balance of the Reductions Escrow is released to the Project Proponent. The amount released can never exceed the total contributed over that Crediting Period.
• Where the reductions Crediting Period is renewed, a new Reductions Escrow begins with the renewed Crediting Period, and any debt still outstanding carries into it and is deducted before further Reduction Certificates are issued.
• Projects must make good any debt outstanding when the Project ends early, or a Crediting Period ends without renewal, with Certificates generated by other projects they operate or Certificates of equivalent quality sourced at their own expense and accepted by Isometric.

Isometric monitors the credited woody-biomass stock for Reversals throughout the removals Crediting Period and the Ongoing Monitoring Period. Monitoring will consist of:

• Annual review of changes in woody biomass cover (e.g., GFC)
• Annual review of changes in vegetation indices

Isometric monitors continuously using satellite burned-area, woody-cover and disturbance products, supplemented by the Project's own field and remote monitoring. Projects must investigate any disturbance that may have reduced the credited woody-biomass stock, act to limit further loss, and implement mitigation to reduce the risk of future Reversals.

Projects must report and investigate any loss of credited woody-biomass stock greater than 1% of the cumulative removals credited to the Project, measured against the total Removal Certificates issued; such losses are compensated in accordance with the Isometric Standard. On detection, the following applies:

• Projects must, within one month of notification, investigate and confirm whether the Reversal has ended and, where it has not, take immediate action to stop or slow the loss and report the near-term adaptive management and mitigation taken or planned for the coming year.
• Projects must, within one year, submit a Reversal report describing the adaptive management and mitigation implemented after the loss. Isometric compiles a Reversal report recording the date, a description, the location and extent of the loss, whether it was avoidable or unavoidable, the quantified loss of woody-biomass carbon, and the impacts on the project activity and the ecosystem, and then initiates the removals Buffer Pool compensation process (see Section 10.3).

A Reversal is quantified as the measured loss of credited woody-biomass carbon stock, converted to CO₂e by the same method used to quantify the removal (see Section 9.3). Additionality is no longer assessed beyond the removals Crediting Period, so the matched control is not used to quantify Reversals during the Ongoing Monitoring Period. Because only living woody biomass is credited, the Protocol conservatively treats the aboveground woody carbon lost in a disturbance as released to the atmosphere in full, with belowground biomass lost in proportion to aboveground biomass. The quantified loss is compensated by cancelling an equal quantity of removals Buffer Pool Credits, up to the liable volume of Removal Certificates (see Section 10.1.1), in accordance with the Isometric Standard. Projects must quantify the remaining woody-biomass stock by field sampling after a Reversal of 20% or more of the cumulative removals credited to the Project.

This section establishes documentation requirements necessary to ensure the accuracy, completeness, and credibility of savanna fire management projects.

Requirements are grouped based on the following:

• Reductions Data
• Removals Data
• Safeguarding Plans

Requirements are further organized into two categories:

• Validation requirements include evidence and documentation required at initial project registration.
• Verification requirements include ongoing evidence to be submitted for each Reporting Period.

Requirements are grounded in the following principles:

• Parameter identification: Quantification of reduction in fire emissions is highly-dependent on accurate and appropriate parameterization, including Vegetation Fuel Type, and fuel class, years since last burning.
• Geographic eligibility: Only projects that operate in ecoregions and vegetation class where emissions factors have been validated are eligible.
• Measurement consistency: The Vegetation Fuel Type map meets a minimum accuracy, and every mapped product is applied identically to baseline and project.
• Appropriate safeguards: Fire management does no harm by maintaining fire-regime biodiversity metrics and including local communities in the planning and execution of the Project.

Validation Requirements

• The pre-project burned-area time series, covering at least ten years or twice the mean fire-return interval, whichever is longer, mapped at 30 m or finer and classified into the early and late dry season by the transition date determined for each year over the Seasonal Observation Area (see Section 4.2.4), used to establish the historical baseline and the fire-return interval.
• The burned-area mapping method and the data sources used, and the documented accuracy of that method or product, from published validation where available.
• The Default Seasonal Cutoff Date adopted (Appendix A, Table A1b), the rainfall dataset and the long-term dry season derived from it, the delineation of the Seasonal Observation Area, and the transition date determined for each historical year.

Verification Requirements

• Annual burned-area maps at 30 m or finer, classified into early and late dry season and stratified by Vegetation Fuel Type, for the project area and the Leakage Monitoring Zone.
• Active-fire products, for example MODIS or VIIRS hotspots, used to cross-check the post-fire burned-area maps.
• The annual dry season and the transition date for each year in the Reporting Period, the number of night-time detections it rests on, and any reclassification of burned area applied under Section 4.2.4.
• The area classified as uncured burns in each year, and the dates of the annual dry season used to classify it.
• The date of every burn in the Reporting Period, the dating precision achieved, and the share of burned area that burnt more than once in the year.
• Total burned area by season, and the proportion of the project area burned, for each year in the Reporting Period, compared with the Historical Reference Period mean.

Validation Requirements

• A Vegetation Fuel Type map validated to at least 80% overall accuracy with ground reference data.
• Evidence that each mapped type meets its eligibility criteria.
• The sample design, including the strata, the mapped area proportion of each class, and the allocation of observations.
• The error matrix in sample counts and in area-weighted proportions, with overall accuracy and each class's user's and producer's accuracy reported with confidence intervals.
• The register of reference observations, recording for each the date, location, observation method and class assigned.

Verification Requirements

• The Vegetation Fuel Type map re-validated at each crediting period rather than annually, since fuel type is structural and slow to change, with any reclassification documented and its accuracy reported.

Validation Requirements

• An initial years-since-last-burnt map built from the running burned-area record of at least five years.

Verification Requirements

• The years-since-last-burnt map updated each year in the Reporting Period from the latest annual burned-area map, setting burnt pixels to age zero and incrementing the rest.

Verification Requirements

• The woody-biomass stock change since project initiation and the matched control's change.
• A review of the control pixels for continued suitability, replacing any that no longer match.
• Through the Ongoing Monitoring Period, continued monitoring of the woody-biomass stock for Reversals, with no new Removal Certificates issued.

Verification Requirements

• Detection and reporting, each Reporting Period and throughout the Ongoing Monitoring Period, of disturbance events and of any measured loss of previously credited woody-biomass stock. A measured loss of credited woody-biomass stock is a Reversal and is compensated from the removals Buffer Pool.

Validation Requirements

• The delineation of the Leakage Monitoring Zone, extending 5 km from the Project Boundary, restricted to eligible fuel and the same Vegetation Fuel Types as the Project and extended across an international border where one is present, and its pre-project historical mean annual fire emissions and their standard deviation.

Verification Requirements

• The Leakage Monitoring Zone is evaluated each Reporting Period, with the same emissions stack, seasonal rule and Vegetation Fuel Types as the project, across both seasons against the threshold of its historical mean plus one standard deviation, and the attributed leakage computed as described in Section 9.4.4. Positive leakage is documented and set to zero.

Validation Requirements

• The fire-management plan, the Stakeholder Engagement Plan and evidence of Free, Prior and Informed Consent, the benefit-sharing agreement, and the grievance mechanism (see Section 6.0).
• The type and extent of pre-project fixed-site productivity in the project area, identifying any areas excluded from the project area and the evidence base for those included (see Section 8.3).

Verification Requirements

• The operational activity data needed to quantify operational emissions, such as fuel used in aerial ignition and in vehicles.
• Updated stakeholder-engagement, grievance, employment and benefit-sharing records.
• A declaration that livestock stocking density in the project area has not increased as a consequence of the project, and, where it has, the extent of the increase and its cause.
• Evidence that fixed-site productivity in the included areas has not been reduced as a consequence of the project (see Section 8.3).

Validation Requirements

• The Project Proponent's committed biodiversity fire-regime values, being a minimum long-unburnt fraction, a fire-frequency cap, and a patch-size or mosaic floor, and their documented scientific basis (see Section 6.3).

Verification Requirements

• The biodiversity metric results computed each Reporting Period.

Projects must reconcile their monitored data each Reporting Period and document the checks in the monitoring report, including:

• the post-fire burned-area maps cross-checked against active-fire products;
• the same resolution, parameters and seasonal classification rule confirmed on the baseline and the project side, with no mixing of resolutions across the baseline-minus-project difference;
• the Vegetation Fuel Type map confirmed against the land-cover product; and
• the Seasonal Observation Area, rainfall dataset and active-fire products confirmed as unchanged from those used for the historical record, or any change documented.

Isometric will re-evaluate the removals Donor Zone for land-cover change each Reporting Period in which removals are quantified, dropping pixels that change land cover, become other carbon projects, or come under a documented early dry season fire-management programme, and excluding the Leakage Monitoring Zone.

This Appendix holds the region-specific default parameters that Section 9.0 applies. A region is eligible only where a validated parameter set appears here (see Section 4.2.1). The Appendix is organized by parameter, so that each variable in Equation 2 maps to one place.

Two regions are provided at first issue. The miombo values are from Russell-Smith et al. 2024⁴ and are stratified at 1000 mm mean annual rainfall, determined from the rainfall dataset in Section 4.2.4. They apply to woody savanna meeting the eligibility criteria in Table A1; no minimum rainfall applies. The northern-Australian values are from Russell-Smith et al. 2009⁴¹ for the high rainfall zone, Yates et al. 2015⁴² for the low rainfall zone and Lynch et al. 2018⁴³ for Pindan, with emission factors from Hurst et al. 1994⁴⁴ and Meyer et al. 2012²⁹. Further regions are added once their parameters are validated by field derivation.

Miombo parameters apply to woodland dominated by Brachystegia, Julbernardia or Isoberlinia, and to Baikiaea plurijuga woodland on Kalahari sands⁴. Woodland (WDL) has woody canopy cover of 20% or more. Open Woodland (OWDL) has woody canopy cover of 10 to 20%. To distinguish it from grassland, it also has at least 10% cover of characteristic miombo or Baikiaea tree and shrub taxa. Closed forest (canopy above 80%) is excluded, because fire does not carry consistently.

Northern-Australian Vegetation Fuel Types are defined by structural formation (canopy height and foliage projected cover) and grass type, within a high-rainfall zone (>1000 mm mean annual wet-season rainfall) and a low-rainfall zone (≤1000 mm mean annual wet-season rainfall).

Miombo rainfall bands are expressed as mean annual rainfall; the northern-Australian bands below are expressed as mean annual wet-season rainfall.

Table A1. Vegetation Fuel Types by region.

Table A1b. Default Seasonal Cutoff Date by region.

Fine and shrub fuels are tabulated by years since last burnt. For fine and shrub fuels these values are the time-since-fire relationship of Equation 3 (see Section 9.2.1.2). When available, coarse and heavy woody fuels use observed values, otherwise they may use pooled means provided here.

Table A2. Miombo fuel load (tonnes of dry matter per hectare) by Vegetation Fuel Type (VFT) and years since last burnt⁴. Pool means for coarse woody debris is 0.55 (±15%), 0.52 (±8%) and 0.39 (±8%), and heavy woody fuel 0.91 (±14%), 0.99 (±11%) and 1.60 (±13%), for OWDL, WDL (≤1000 mm) and WDL (>1000 mm). Late dry season fine-fuel load is adjusted upward for litter accession. Litter is 60%, 67% and 76% of the annual fine-fuel mean in the three types. Values in brackets are relative standard errors of the mean used in the uncertainty propagation (Section 9.5).

Table A3. Northern Australian fine-fuel load (tonnes of dry matter per hectare) by Vegetation Fuel Type (VFT) and years since last burnt. Pindan (low zone, no seasonal difference) accumulates over a longer interval, reaching, at years 1 to 10 or more: 2.27, 3.17, 3.85, 4.41, 4.91, 5.36, 5.77, 6.15, 6.50, 6.84. The woodland, mixed grass (high zone) curve carries a relative error of ±10%, from the standard errors of the fuel-accumulation coefficients for eucalypt woodland in Russell-Smith et al. 2009, Table 5⁴¹, used in the uncertainty propagation (Section 9.5).