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 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 a decrease of 10 to 15% in combustion of fine fuels when burning in the early season rather than the late. Furthermore, these same early season fires in miombo regions burn approximately a third less total area compared to late season fires burn. In total, the per-unit-area early season emissions are estimated to be roughly 34 to 56% of late season levels⁴ per unit area. 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;
• baselines are dynamic and conservative, separating project performance from regional fire-weather variation;
• 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
• 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:

• Savanna Fire Management, consultation draft (CN0135), Verra, 18 February 2026
• 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 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 severity⁸;
• Including emission factors for carbon monoxide (CO)⁹;
• Broadening the eligible regions and Vegetation Fuel Types as region-specific emission factors and fuel parameters are validated;
• Validating fuel accumulation with field measurements;
• Incorporating the effects of varying levels of herbivory¹⁰.

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.

To qualify for crediting under this Protocol, a Project must meet all of the conditions set out in this section.

The geographic Project Boundary must encompass all areas in which the Project Proponent conducts savanna fire management for crediting. It may be a single continuous area or a collection of discrete management units. The Project Boundary must be fixed at project initiation and may be changed only by the addition of new areas once the Crediting Period has begun.

Areas of ineligible vegetation within the Project Boundary must be mapped and excluded from quantification. Land is ineligible where:

• it is not one of the Vegetation Fuel Types defined for the project's region in Appendix A;
• where it does not meet the eligibility thresholds set in Appendix A;
• where it falls within the ineligible project types listed in Section 4.3; or
• where it is not vegetated, such as water bodies, roads, settlements and cleared land.

Ineligible land is excluded from quantification for both the baseline and the Project.

A Project must manage fire so as to reduce fire emissions, increase woody-biomass carbon, or both, through one or more of the following activities:

• shifting burning from the late dry season to the early dry season; and/or
• reducing the total area burned and the frequency of fire, including through patchy early dry season burning that fragments the spread of late dry season fire.

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-offs¹¹.

Projects must distinguish between the early and late dry season by a Seasonal Cutoff Date derived from observed fire behavior in the Donor Zone (see Section 4.2.4). The Seasonal Cutoff Date is fixed before the Crediting Period begins and is known to the Project in advance of each burning season, and is corrected after the fact where fire behavior transitions materially earlier than expected.

• The same derivation must be applied to both the baseline record and the project period, and both must be classified from transition dates determined over the same Donor Zone.

This Protocol applies to fire-prone woody savannas, typically with woody canopy cover >10%¹².

Project areas must meet all of the following conditions:

• it is woody savanna, confirmed by intersection with a published savanna ecosystem boundary;
• the dry season is at least four months long, defined as a period in which cumulative rainfall is less than 10% of mean annual wet-season rainfall; and
• it lies between 23.5°N and 23.5°S.

Emission factors and fuel parameters are region-specific and cannot be generalized across regions of the world or other vegetation types^13,14,15. A region is therefore eligible only where a validated default parameter set is provided for it in Appendix A. Further regions, and further Vegetation Fuel Types, become eligible once their parameters are validated by field studies and in consultation with Isometric.

The Project Boundary must be stratified into the Vegetation Fuel Types defined for the project's region in Appendix A.

A Vegetation Fuel Type is a mapped class of structural formation, canopy cover, and grass type, and each carries its own fuel-load, 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 Vegetation Fuel Type map must be validated against reference observations that are independent of the data used to produce it. Validation must use probability-based sampling stratified by mapped class, including the ineligible class, and must achieve at least 80% overall accuracy, with the full confusion matrix reported, area-weighted, and with the user's and producer's accuracy of each class reported with its confidence interval¹⁶.

Reference observations may be field plots, higher-resolution imagery, or systematic aerial observation, provided that each determines the class of the assessment unit against the same class definition used to produce the map. Where a distinction between classes depends on an attribute that cannot be reliably determined by the observation method used, such as the composition of the understory, reference observations for that distinction must be made in the field.

An observation matches where the class occupying the greatest proportion of the assessment unit is the same as the reference class assigned to it. A match must not be recorded on the basis that the reference class occurs somewhere within the assessment unit.

The number of reference observations must be sufficient to achieve a stated precision on the overall accuracy, and each mapped class must carry enough observations for its own accuracy to be estimated. Where a class cannot be assessed, it must be merged with an ecologically similar Vegetation Fuel Type, taking the more conservative of the two parameter sets, or classified as ineligible. A class must not be retained in the map, and in the emissions calculation, without a validated accuracy estimate.

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 and re-scaled each Reporting Period for regional fire-weather variation. To be eligible, a Project must be able to establish this baseline before project activities begin:

• the project area must have a fire regime that is predominantly late dry season, evidenced from a burned-area record covering at least ten years, or at least twice the mean historical fire-return interval, whichever is longer;
• the baseline must be fixed before project activities begin, and is not reassessed after the project starts; and
• a regional donor area must be available whose historical fire regime tracks the project area's historical fire regime.

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.

For each year, the dry season is identified from that year's rainfall record, beginning on the day before the first four-week period with no significant rainfall and ending at the onset of the wettest eight-week period.

The early/late boundary is the Seasonal Cutoff Date, a single date fixed before the Crediting Period begins. Burned area is assigned to the early or late dry season by its observation date relative to the Seasonal Cutoff Date, subject to the adjustment set out below.

The Seasonal Cutoff Date is derived from the transition between early and late dry season fire behavior observed in the Donor Zone described in Section 9.2.2.1. For a given year, the transition date is the date on which night-time fire begins to persist, determined from day-time and night-time active-fire detections over the Donor Zone as follows¹⁷:

• cumulative day-time and night-time active-fire counts are computed across the year;
• the gradient of the ratio of night-time to day-time cumulative counts is calculated;
• periods of night-time fire activity separated by seven or more consecutive days with no detection are treated as separate activity windows, and any window of seven days or less is disregarded; and
• the transition date is the date of the maximum gradient within the first remaining activity window.

The Seasonal Cutoff Date is the mean of the transition dates across the Historical Reference Period, being twice the mean fire-return interval or ten years, whichever is longer (see Section 9.2.2). It is determined at Validation and fixed for the Crediting Period.

The transition date is determined over the Donor Zone, since it reflects the environmental conditions and should not be confounded by the Project’s activities. A Project that succeeds in suppressing late dry season fire would move a project-area transition date later, and so weaken the test intended to detect that suppression. The Donor Zone is exogenous to the project activity, is already delineated for the dynamic adjustment (see Section 9.2.2.1), and is large enough to yield the active-fire detections the derivation requires.

Each year of the historical burned-area record is classified using the transition date determined for that year, rather than the Seasonal Cutoff Date.

Burned area in each Reporting Period is classified using the Seasonal Cutoff Date, subject to the following adjustment. Where the transition date for that year falls more than ten days earlier than the Seasonal Cutoff Date, all burned area observed between the transition date and the Seasonal Cutoff Date must be reclassified to the late dry season.

• No adjustment is made where the transition date falls later than the Seasonal Cutoff Date, or where it falls earlier by ten days or less.
• Where the transition date cannot be determined for a year, the Seasonal Cutoff Date applies without adjustment.

Fires are not expected outside the dry season, but where they are detected they must still be counted and classified by the same division. A fire detected after the late dry season ends and before the midpoint of the wet season is assigned to the late dry season, and a fire detected after the wet-season midpoint and before the early dry season begins is assigned to the early dry season. Off-season fires are therefore treated consistently on the baseline and the project side. Future versions will seek to use remotely sensed fire severity (see Section 3.0).

The following are not eligible under this Protocol:

• non-woody grassland (<10% woody canopy cover), where low-intensity early dry season fire can promote woody thickening and encroachment;
• regions and Vegetation Fuel Types for which region-specific emission factors and fuel parameters have not been validated;
• project activities that depend on increasing livestock stocking density; and
• vegetation clearing interventions, and any activity that depends on clearing.

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.

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.

At the end of the initial Crediting Period, the removals pool enters a mandatory Monitoring Period of 50 years. No further Removal Credits are issued during the Monitoring Period. Its purpose is to confirm the permanence of the woody-biomass removals already credited.

A measured 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 reductions pool may be renewed and continue to credit emission reductions in parallel with the removals 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.

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. At each renewal, the Project must re-confirm that it continues to meet the eligibility conditions of Section 4.0 and that it still holds 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. Credits are issued for each verified Reporting Period.

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 Buffer Pool. 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 Credits may be issued annually. Crediting is not conditional on the Project reaching a steady fire regime.

The Project must identify, assess and manage the environmental and social risks of its fire-management activities throughout the life of the Project. 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.

An environmental and social risk assessment must be completed in accordance with the Isometric Standard, identifying risks to local communities, customary land users and ecosystems, followed by mitigation plans and a monitoring plan that detects adverse effects and allows activities to be adapted or paused. 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 leakage¹⁸.

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

Project Proponents must identify every group that uses fire in, or depends on, the landscape within the Project Boundary and its surroundings, including pastoralists, hunters, farmers, resource harvesters, and Indigenous and customary communities.

• Communities surrounding the Project but not directly involved, should be informed of the Project and any grievance resolution mechanisms to address unintended fire or other spillover consequences.
• Where the Project abuts an international border, groups on the other side of the border who use the same landscape must be identified and engaged a similar basis.

The Project Proponent must assess whether project activities will affect Indigenous Peoples or customary land users, including any effect on land or territories they own, occupy or use, and on the natural resources their livelihoods or cultural practices depend on, consulting an independent expert where appropriate. Where such impacts are identified, the Project Proponent must obtain Free, Prior and Informed Consent (FPIC), consistent with the United Nations Declaration on the Rights of Indigenous Peoples:

• Free: engagement is free of intimidation, coercion and manipulation.
• Prior: consent is sought early, before project activities begin, on a timeline informed by customary practice.
• Informed: information is provided in an accessible language and medium, presented fairly, and includes the value of the Credits, the expected revenue, the benefit-sharing arrangement, and the alternatives to participation.
• Consent: consent is freely given, may be made conditional, and may be withdrawn.

Project Proponents must be able to evidence that engagement met these principles, including:

• the measures taken to reach all groups and the share of adults reached;
• the medium and language used; and
• how stakeholder input was incorporated into the fire-management design.

The VVB may interview or survey stakeholders to confirm this. Records of the FPIC process and the benefit-sharing agreement must be made publicly available, with personal information removed.

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.

Project Proponents must describe how they include local communities in their project design.

This includes:

• recognize and use Indigenous and traditional fire knowledge in the design and conduct of the fire regime;
• employ and resource local and Indigenous communities, such as through a community fire-ranger model^19,20; and
• build on customary fire institutions rather than replacing them, combining customary and contemporary fire management, an approach shown to deliver both emission reductions and livelihood benefits²¹.

Benefits amongst Project participants must be shared equitably and in a decentralized way, with an explicit guard against centralizing carbon revenue or sidelining Indigenous leadership, which may lead to the failure of savanna fire management schemes²².

Project Proponents must disclose any benefit-sharing arrangement,

Benefit-sharing agreements should:

• disclose the share of revenue due to participating communities agreed to through the stakeholder engagement process (see Section 6.2.1); and
• favor local employment and reach women and other groups who might otherwise be excluded.

The Project Proponent must:

• describe how it protects human rights and applies anti-discrimination policies, including on the basis of gender;
• establish a grievance mechanism through which community members can raise grievances, feedback and complaints, with independent mediation and a documented resolution process; and
• describe its hiring practices and the short, medium and long-term employment created, and the share recruited from the local community.

Early dry season burning pursued without regard to fire-regime diversity can homogenize habitat and harm fire-sensitive species¹⁷.

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

Targets are determined by Project Proponent and should be scientifically supported.

Examples of do-no-harm commitments²³ include but are not limited to:

• a minimum proportion of vegetation left long-unburnt;
• a maximum on fire frequency; and/or
• a maximum fire patch-size or other fragmentation metric²⁴.

The commitment is monitored from the fire-regime data the Project already produces (season, frequency, time since last burnt, and patch size), and a continuous breach is a safeguard failure requiring monitoring in consultation with Isometric.

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.

The Project Proponent must document the Project in a Project Design Document (PDD), as required by the Isometric Standard. For a savanna fire management Project, the PDD must include:

• 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, and the donor area used to adjust it;
• 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 ex-ante estimate of the reductions and removals expected over the crediting period; and
• the monitoring plan for fire, woody biomass, leakage and reversals.

A Project must be validated, and its net reductions and removals verified, by an independent Validation and Verification Body (VVB), consistent with this Protocol and 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.
• 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.

Project Validation and Verification must incorporate site visits to the project area, in accordance with the requirements of ISO 14064-3, 6.1.4.2, and the Isometric Standard. This must include, at a minimum, a site visit during Validation and during the initial Verification, where the validator observes fire-management operations and the field reference data used to validate the burned-area and Vegetation Fuel Type maps. Given the scale and remoteness of many savanna projects, the site visit may be combined with the ground-truthing of those maps.

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.

Verifiers and validators must comply with the requirements defined in the Isometric Standard. In addition, VVB teams must 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. Where multiple parties are involved, and to avoid the double counting of net reductions or removals, a single Project Proponent must be specified contractually as the sole owner of the Credits. Contracts must comply with all requirements defined in the Isometric Standard.

The Project Proponent must also hold the legal right to implement the project activity.

• For savanna fire management, this includes any permit or exemption needed to conduct prescribed early dry season burning where a fire-exclusion law applies²⁵.

The Project Proponent must demonstrate additionality through compliance with the Isometric Standard. The counterfactual scenarios and baselines used to assess additionality must be project-specific and are described in Section 8.0. For emission reductions, the dynamic-adjusted baseline of the project area's own fire emissions, while removals use a matched-control baseline of woody-biomass change in a comparable control area. Both counterfactuals are dynamic, so that regional fire-weather variation and regional woody-biomass trends are separated from project performance.

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; Credits issued for current or past Crediting Periods are not affected.

Common practice analysis tests whether the project activity is already widespread in the region without carbon finance, since an activity that is common practice is not additional. The following steps must be taken to demonstrate that, without Carbon Finance, the project activity is not common practice, in accordance with the requirements defined in the Isometric Standard.

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^26,27. The project activity is the deliberate management of fire that shifts burning to the early dry season and reduces the total area burned in late dry season fires. The analysis must therefore measure how widespread that managed regime is, not how widespread fire is.

1. Define the project activity as managed early dry season fire that shifts burning from the late to the early dry season and reduces total burned area and late dry season fire frequency, as set out in Section 4.1.
2. Identify the applicable geographic area, as described in the Isometric Standard. The regional donor area used for the baseline (see Section 8.2) is a suitable basis, since it is already delineated as comparable land outside the Project, of the same vegetation and landcover class.
3. Identify and explain any essential distinctions between the Project and similar activities, as described in the Isometric Standard. For savanna fire management the essential distinction is between a deliberately managed early dry season regime and fire that merely happens to occur early for reasons unrelated to management.
4. Assess the market penetration rate as the area proportion of the comparable class already under a managed early dry season regime without Carbon Finance, using either a survey-based approach or relevant data from existing literature:

• Survey-based approach: survey a representative sample of similar land managers from within the geographic area within five years of the Project start date, and calculate the cumulative market penetration rate, as an area percentage, of the managed regime among managers who have not received Carbon Finance revenue.
• Data from existing literature: a burned-area record classified into the early and late dry season by the same rule the Project applies (see Section 4.2.4), supported where available by evidence of fire-management programs, ranger activity or fire-management institutions, derived from data collected within five years of the Project start date, relevant to the project area, conservative in that it does not distinguish managers incentivized by Carbon Finance from those who are not, and publicly available as government or census data, peer-reviewed scientific literature, or transparent independent research or reports.

In accordance with the Isometric Standard, the project activity demonstrates common practice additionality where the market penetration rate is at or below 20% of areas conducting early dry season fire management without Carbon Finance.

The uncertainty in the overall estimate of the net CO₂e impact of the Project must be accounted for, and the net impact for each Reporting Period must be conservatively determined, in accordance with the Isometric Standard. The quantity credited is the 16th-percentile lower bound of the net estimate, about the mean minus one standard deviation, after propagating the Project's measurement uncertainties. The error-propagation method is set out in Section 9.5.

The Project must report a list of the key variables used in the net reduction and net removal calculations and their individual uncertainties, together with a description of the uncertainty-analysis approach, including:

• the burned-area mapping omission and commission rates, from the accuracy assessment;
• the fuel-load accumulation curves, combustion-completeness and emission-factor parameters, with their sample sizes and errors, and the global warming potentials applied;
• the patchiness factors and the donor adjustment; and
• the woody-biomass measurements and the matched-control comparison.

A sensitivity analysis that shows the effect of each input parameter's uncertainty on the final net estimate must be provided, in enough detail that a third party can 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 and their accuracy assessment;
• 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)²⁸, 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.

A cradle-to-grave GHG Statement must be prepared encompassing the GHG emissions relating to the activities outlined 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 removals account, as set out in Section 8.3. Emissions 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. The Project Proponent is responsible for identifying 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 regrowth²⁹. 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^30,31

The system boundary must include all relevant GHG SSRs controlled by and related to the Project, including but not limited to the SSRs set out in Table 1. If any GHG SSRs within Table 1 are deemed not appropriate to include in the system boundary, they may be excluded, provided that robust justification and appropriate evidence are provided 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.

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

• consider all GHGs associated with its 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 two dynamic baselines, one for each account. Each baseline instead carries a dynamic component, derived from areas outside the project, that tracks regional conditions through the Crediting Period.

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 historical fire emissions, re-scaled each Reporting Period for regional fire conditions. It has two structural components.

The first component is a static historical baseline. The project area's historical woodland fire emissions, computed over a fixed pre-project window and frozen at project start.

The second component is a dynamic adjustment: a single dimensionless modifier that re-scales the historical baseline fire emissions each Reporting Period by how far regional fire conditions differ from their historical norm. This weather-stripping step separates climate-driven variation from the Project’s performance. The process for calculating the reductions baseline is described in Section 9.2.2.

Negative annual results, where the Project's measured emissions exceed its adjusted baseline emissions, are considered Reversals and must be compensated from the reductions Buffer Pool (see Section 10.3). If measured emissions exceed the reductions Buffer Pool amount, any residual emissions shall be carried over to the next reporting period and deducted from future reductions.

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 the same Donor Zone used in the reductions baseline. 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 belt, 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^32,33, 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 benefits³⁴.

The project does not, however, 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. The residual risk is activity-shifting displacement, which the Protocol monitors.

This Protocol requires a Leakage Monitoring Zone with no de minimis exemption. The Leakage Monitoring Zone extends at least 10 km outward from the Project Boundary regardless of international boundaries, and its width is increased where necessary so that the Leakage Monitoring Zone area is at least equal to the project area. 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 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.

The comparison follows the same logic as the dynamic adjustment (Section 9.2.2.1): measured emissions are set against a historical mean that has itself been re-scaled for regional fire conditions, so that only the departure from those conditions is attributed to the Project. This monitoring of the Leakage Monitoring Zone is deliberately distinct from the coarse relative index used for determining the counterfactuals for reductions (Section 9.2.2.1) or removals (Section 9.3.2.1).

Figure 1. Schematic of project area, including the Donor Zone and Leakage Monitoring Zone.

Emissions displacement attributed to the Project is deducted from the Removal Credits (see Section 9.4.4). The attributed emissions increase is the rise in emissions above the historical baseline, after applying the same dynamic adjustment to account for changing conditions. The attributed leakage, together with operational emissions, is deducted from the Reporting Period's removals first, as described in Section 9.0. 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 ignition³⁵.

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 Credits. A Reduction Credit represents one tonne of CO₂e of reduced methane and nitrous oxide emissions. A Removal Credit 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 dynamic-adjusted fire emissions baseline (see Section 8.2.1), and the removals counterfactual is a matched control of changing biomass within the region (see Section 8.2.3).

The project's life-cycle emissions and leakage are the only deductions shared between the two accounts; together they form the project emissions term (Equation 13), which is deducted from the removals account first and reduces the reductions account only to the extent it exceeds the Reporting Period's removals. Because it is measured as methane and nitrous oxide from displaced fire, it is deducted from the reductions account alone (see Section 9.4.4).

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

The reductions account credits 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 dynamic-adjusted baseline fire emissions and the project's measured fire emissions::

$CO_2e_{Reductions,RP} = CO_2e_{Fire,Baseline,RP} - CO_2e_{Fire,Project,RP}$
(Equation 1)

Where:

• $CO_2e_{Reductions,RP}$ represents the gross emission reductions, as CH₄ and N₂O, for the Reporting Period, $RP$, in tonnes of CO₂e.
• $CO_2e_{Fire,Baseline,RP}$ represents the dynamic-adjusted 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.

The Project's life-cycle emissions and leakage 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 debit drawn from the reductions Buffer Pool.

The project's fire emissions for the Reporting Period, $CO_2e_{Fire,Project,RP}$, and the unadjusted baseline's fire emissions, $CO_2e_{Fire,Historical,RP}$ (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:

$CO_2e_{Fire}=\sum\limits_s\sum\limits_v\sum\limits_f(A_{v,s} \times P_{v,s} \times FL_{v,f} \times CC_{s,f} \times \sum\limits_g(EF_{g,f} \times GWP_g))$
(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).
• $GWP_g$ represents the 100-year global warming potential for gas $g$, from IPCC AR6³⁶: 27 for CH₄ and 273 for N₂O.

Where an emission factor is reported as a fraction of fuel carbon or nitrogen, as the regional parameter sets in Section 12.5 are, it must be converted 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.

Burned area must be mapped at 30 m or finer, classified by season and stratified by Vegetation Fuel Type, on both the baseline and the project side (see Section 8.0). 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. Where the assignment of small burnt patches is ambiguous, they must be conservatively 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.
• $a(f)$ 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.

The years-since-last-burnt map must be updated 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 same measured 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^37,38,39; the seasonal contrast is carried by combustion completeness and patchiness, not by the emission factor.

The Project’s counterfactual baseline fire emissions, $CO_2e_{Fire,Baseline,RP}$, represent the estimate of fire emissions in the project area if the fire regime had not been altered. Because the precise quantity of these hypothetical emissions is unknowable, this value is estimated using the project area’s historical emissions dynamically adjusted for changing climate and weather influences as observed from the broader region.

The dynamic-adjustment re-scales the historical baseline as follows:

$CO_2e_{Fire,Baseline,RP}=CO_2e_{Fire,Historical,RP} \times DA_{RP}$
(Equation 4)

Where:

• $CO_2e_{Fire,Historical,RP}$ represents the project area's pre-project average historical annual fire emissions, in tonnes of CO₂e, and is calculated using the same equations as set out in Section 9.2.1. This emissions value is calculated once at the beginning of the Project and set for the entire Crediting Period.
• $DA_{RP}$ represents the dimensionless dynamic adjustment calculated for each Reporting Period, $RP$ (see Section 9.2.2.1). $DA_{RP}$ equals one when regional fire activity is at its pre-project norm, rises above one in regionally high-fire periods, and falls below one in quiet ones.

Projects must calculate the average historical annual baseline over twice the mean fire-return interval or 10 years⁴⁰, whichever is longer.

• The mean fire-return interval is defined as the average time between successive fires within the project area. This must be determined through:
  • remote sensing data; or
  • scientific literature appropriate to the ecoregion.
• The window should be long enough to represent the project's typical fire emissions.

The dynamic adjustment is the primary weather-stripping mechanism for determining the counterfactual fire emissions. Changes in regional fire conditions are evaluated by observing fire within a nearby Donor Zone. The Donor Zone is a regional annulus around the Project Boundary, restricted to the same landcover class as the Project excluding a Leakage Monitoring Zone and any other carbon projects with managed fire regimes. This Donor Zone is the same as the area used for estimating the counterfactual for removals (see Section 9.3.2.1).

The dynamic adjustment factor is calculated as the Donor Zone's recent burned area relative to its historical burned area:

$DA_{RP}=\frac{BA_{Donor,RP}}{BA_{Donor,Historical}}$
(Equation 5)

Where:

• $BA_{Donor,RP}$ represents the donor area's burned area fraction in the credited landcover class, averaged over the trailing dynamic-adjustment window ending in the Reporting Period, $RP$, in hectares.
• $BA_{Donor,Historical}$ represents the donor area's burned area fraction in the same landcover class averaged over the same trailing window length ending at Project initiation, in hectares.

The length of the trailing window for calculating the burn-area fraction must be sized so that the Donor Zone's historical fire regime tracks the project area's historical fire regime.

The steps to determine the window size, $W$, are as follows:

1. Build two annual time series over the longest clean pre-project record available: the Donor Zone's burned-area fraction and the project area's burned-area fraction, both in the credited vegetation class.
   • For this relative index the Donor Zone burned area may be taken from coarse satellite data, e.g. FireCCI51. Project emissions are quantified separately at finer resolution.
2. For a candidate window length $W$, smooth both series with a $W$-year trailing mean.
3. Compute the squared Pearson correlation (r²) between the two smoothed series.
4. Adopt the shortest $W$ for which r² ≥ 0.50. Annual fire is idiosyncratic, so the unsmoothed series correlate only weakly; averaging over a trailing window suppresses that noise and exposes the shared regional climate signal, so r² rises with $W$. Taking the shortest qualifying $W$ keeps overlapping windows from inflating the apparent agreement and preserves independent information in the signal.
5. A window length is admissible only if the pre-project record spans at least three non-overlapping windows (record length ≥ $3W$). If the shortest $W$ reaching r² ≥ 0.50 is not admissible, the Donor Zone fails this test.
6. If no Donor Zone geometry and admissible window length reaches r² ≥ 0.50, the Project Proponent must consult Isometric to re-delineate the Donor Zone (geometry, distance bands, or vegetation restriction).

The dynamic adjustment window is kept short to limit autocorrelation between overlapping windows, and so need not match the longer historical-baseline window.

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).

Gross removals for the Reporting Period are the project's woody-biomass stock change above the matched-control counterfactual:

$CO_2e_{Removals,RP} = CO_2e_{Biomass,Stored,RP} - CO_2e_{Biomass,Counterfactual,RP} - CO_2e_{Emissions,RP}$
(Equation 6)

Where:

• $CO_2e_{Removals,RP}$ represents the net carbon removals in biomass for the Reporting Period, $RP$, in tonnes of CO₂e.
• $CO_2e_{Biomass,Stored,RP}$ represents the change in the project's woody-biomass carbon stock over the Reporting Period, $RP$, expressed as CO₂e.
• $CO_2e_{Biomass,Counterfactual,RP}$ represents the woody-biomass stock change that would have occurred without the project, determined from the matched control (see Section 9.3.2)
• $CO_2e_{Emissions,RP}$ represents the project's life-cycle emissions and leakage for the Reporting Period, $RP$ (see Section 9.4)

The woody biomass carbon stock change over the Reporting Period is:

$CO_2e_{Biomass,Stored,RP}=\Delta M_{Biomass} \times CF \times \frac{44}{12}$
(Equation 7)

Where:

• $\Delta M_{Biomass}$ represents the change in total woody biomass over the Reporting Period, $RP$, 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.

The carbon fraction, $CF$, must be chosen from the following hierarchy:

1. A regional and species-specific factor that is justified based on scientific literature (e.g., Doraisami et al., 2022⁴¹). 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^42,43, can be used.

The total carbon stored in woody biomass over a Reporting Period is calculated by taking the difference between the start and end of the Reporting Period. At each time point, the biomass is calculated as:

$M_{Biomass}=M_{AGB}+M_{BGB}$
(Equation 8)

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 9)

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.

If sufficient evidence is provided to demonstrate that no suitable project-specific factor can be obtained, matching to the ecological zone and continent of the project area, based on the IPCC 2019 Table 4.4⁴⁴, must be used. In this case, sufficient evidence documenting the unsuccessful search for project specific factors must also be supplied. Acceptable evidence must show:

• 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 Database⁴⁵.

The uncertainty in selected RS factors must be reported from the same source dataset. For example, the IPCC 2019 Chapter on Forest Land⁴⁶ 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. Project Proponents must describe in the PDD which option is used, and adhere to the requirements of that approach. 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 Credits 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^47,48,49,50.

Dynamic baselines will be independently determined and transparently reported by Isometric at each Verification to determine any deduction in Credit issuance based on the Baseline scenario. Credit 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.

The zone from which control pixels will be selected, termed the Donor Zone, must meet 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 primarily early dry season managed fires; and
• does not contain unrepresentative land use types (e.g., urban areas); and
• be subject to the same or similar relevant regulations, government incentives, and programs as the project area; and
• must demonstrate 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. This Donor Zone does not need to be the same area as that used in Section 9.2.2.1 for determining the relative amount of fire for the dynamic adjustment of the historical fire emissions.

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. These layers must cover the entire project area and Donor Zone at the same resolution for at least five historical time points relative to the start of the Project. Each historical time point must be separated by at least 1 year.

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

• The proxy must be 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), it will be excluded⁵¹. 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^52,53.
• 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 itself is used instead of a proxy map, then the AGB map product used to generate the dynamic baseline must 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 must 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 over the Reporting Period 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:

$\Delta C_{proxy}=C_{proxy}(t_2)-C_{proxy}(t_1)$

(Equation 10)

Where:

• $t_1$ and $t_2$ are the start and end of the Reporting Period, $RP$, respectively.
• $\Delta C_{proxy}$ is the change in carbon proxy value over the Reporting Period, $RP$.

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 differences of the proxy change between the project and control pixels:

$$CO_2e_{Biomass,Counterfactual,RP}=CO_2e_{Biomass,Stored,RP} \times \frac {\mu \Delta C_{proxy,project}-\mu \Delta C_{proxy,difference}}{\mu \Delta C_{proxy,project}}$$

(Equation 11)

Where:

• $CO_2e_{Biomass,Counterfactual,RP}$ is the total counterfactual CO₂ removed from the atmosphere and stored as organic carbon in living woody biomass in the absence of project activities for the Reporting Period, $RP$, in tonnes of CO₂e.
• $CO_2e_{Biomass,Stored,RP}$ is the CO₂e stored for the Reporting Period, $RP$, as calculated in Section 9.3.1.
• $\mu \Delta C_{proxy,project}$ is the mean change in carbon proxy over the Reporting Period for all project pixels.
• $\mu \Delta C_{proxy,difference}$ is the mean difference in proxy change between each project pixel and its matched control pixels. (e.g., $\Delta C_{proxy,project,i}-\mu \Delta C_{proxy,control,i}$)
  • $\mu \Delta C_{proxy,control,i}$ is the average change in forest carbon proxy over the Reporting Period, $RP$, in the group of control pixels matched to project pixel $i$.
  • $\Delta C_{proxy,project,i}$ is the change in forest carbon proxy over the Reporting Period, $RP$, in project pixel $i$.
  • $i$ denotes each individual project pixel.

To be considered additional, the change in proxy value in the project area, $\mu \Delta C_{proxy,project}$, must be statistically greater (p < 0.05, inclusive of uncertainty) than the mean change in proxy value for the matched control pixels, $\mu \Delta C_{proxy,project}$. If the mean proxy change in the control pixels is negative such that the resulting product of Equation 18 is negative, the counterfactual carbon storage, $CO_2e_{Biomass,Counterfactual,RP}$, will be assumed to be 0 in order to ensure the accounting of carbon storage is limited to removals.

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

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

(Equation 12)

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 within control plots as outlined in Section 9.3.2.1. In the event that control pixel matches are no longer suitable, replacements will be selected for the impacted project pixels.

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

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

(Equation 13)

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.
• $CO_2e_{Leakage,RP}$ represents the attributed leakage for the Reporting Period, $RP$, in tonnes of CO₂e (see Section 9.4.4).

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 v1.1.

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. They must be attributed 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.

$CO_2e_{End-of-Life,RP}$ ​must be estimated upfront and allocated in the same way as set out for calculation of $CO_2e_{Establishment,RP}$.

Given the uncertain nature of $CO_2e_{End-of-Life,RP}$ emissions, assumptions must be revisited at each Reporting Period and any necessary adjustments made.

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. The Leakage Monitoring Zone's pre-project historical emissions are re-scaled by the same dynamic adjustment used for the Project’s baseline, so that a regionally wet or dry year is not misread as leakage :

$CO_2e_{Leakage,RP}=CO_2e_{LeakageMonitoringZone,RP}-(CO_2e_{LeakageMonitoringZone,Historical} \times DA_{RP})$
(Equation 14)

Where:

• $CO_2e_{LeakageMonitoringZone,RP}$ represents the measured fire emissions in the Leakage Monitoring Zone during the Reporting Period, $RP$, in tonnes of CO₂e.
• $CO_2e_{LeakageMonitoringZone,Historical}$ represents the Leakage Monitoring Zone’s pre-project historical mean annual fire emissions, in tonnes of CO₂e.
• $DA_{RP}$ represents the dynamic adjustment for the Reporting Period, $RP$, as in Equation 5, computed once for the Project from the Donor Zone and applied here unchanged .

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.

Any positive value of the attributed leakage is deducted in full. There is no de minimis exemption and no threshold below which an increase in the Leakage Monitoring Zone is disregarded.

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

GHG emissions accounting must follow 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.

Energy-related emissions, such as fuel for aerial ignition and vehicles and electricity for monitoring equipment, must be calculated 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. In accordance with the Isometric Standard, the credited quantity is the 16^th-percentile lower bound of the net estimate, about the mean minus one standard deviation, 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 omission and commission, from the reported accuracy assessment;
• 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 dynamic counterfactuals.

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. It nonetheless contributes to a separate reductions Buffer Pool, which compensates reductions debits where project emissions exceed the baseline, as set out in Section 10.3.

The durability of a Removal Credit is the length of the removals Monitoring Period, and therefore 50 years following the end of the initial Crediting Period (see Section 5.0). No further Removal Credits are issued during the Monitoring Period; throughout it the Project must maintain the credited woody-biomass stock and monitor it for loss.

The duration of the Ongoing Monitoring Period must not exceed any of the following:

• Land tenure and legal right. Project Proponents must hold and maintain access to the project area and the legal right to manage fire on it, including any permit or exemption required under a fire-exclusion law, throughout the Monitoring Period (see Section 7.3).
• Financial continuity. Project Proponents must demonstrate continued funding or financial incentive to maintain the woody-biomass stock throughout the Monitoring Period, since the stock is held in place only while the managed fire regime continues.
• Continued fire management. Project Proponents must continue the early dry season burning and risk-mitigation activities that maintain the stock and prevent severe late dry season fire from releasing it throughout the Monitoring Period.

Two savanna-fire-management-wide Buffer Pools managed by Isometric, one for reductions and one for removals, insure the Credits against Reversals. Throughout the Monitoring Period, Isometric monitors for Reversals to ensure the Credits achieve their stated durability, and on detection and quantification of a loss an equal quantity of removals Buffer Pool Credits is cancelled (see Section 10.3 and Section 10.4).

A long-term durability plan is required to maintain the woody-biomass stock beyond the Monitoring Period. The plan may rest on, and ideally combine, the following:

• ongoing financial sustainability beyond the Monitoring Period, for example through conservation finance or a transition to alternative income streams;
• a plan to secure legal protection of the project area beyond the Monitoring Period; and
• building local technical capacity or employment to sustain long-term management of the woody-biomass stock.

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 disturbance⁵⁴. 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, and rising ignition pressure in the surrounding landscape. 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. A sustained early dry season burning that fragments the landscape and prevents severe late dry season fire from carrying through the woody stock. Beyond this, the Project must maintain firebreaks and fire-suppression capacity, plan for drought, sustain the community fire-governance arrangements that keep fire users integrated rather than displaced, and identify and mitigate risks specific to its region, such as cyclone exposure or regionally pervasive pests.

Two Buffer Pools managed by Isometric and governed by the Isometric Standard insure the Project's Credits: a removals Buffer Pool, which insures the Project's credited removals against Reversals, and a reductions Buffer Pool, which compensates the reductions account when project emissions exceed the baseline. The two Buffer Pools are held and drawn on separately, and a debit on one account is not compensated from the other Buffer Pool. On detection and quantification of a loss, an equal quantity of Credits is cancelled from the corresponding Buffer Pool, so that the net impact already credited remains whole.

Each Reporting Period, the Project contributes a flat 20% of the Removal Credits generated that period to the removals Buffer Pool. The Project also contributes a flat 5% of the Reduction Credits generated that period to the reductions Buffer Pool. Both contributions are fixed proportions and are not risk-rated; neither rate varies with the reversal-risk rating required under Section 10.2 or with any change in the Project's risk profile over the Crediting or Monitoring Period. The reductions account draws on the reductions Buffer Pool for the debits described below.

The contributions are held in two savanna-fire-management-wide Buffer Pools managed by Isometric, one for reductions and one for removals. The composition of each 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 initial Crediting Period is incorporated into the removals quantification at each Verification. Where the net removal for a Reporting Period is negative, an equal quantity of removals Buffer Pool Credits is cancelled.
• A Reversal of removals during the removals Monitoring Period is quantified as described in Section 10.4 and fully compensated from the removals Buffer Pool within one year of the loss event.
• A reductions debit, where the project area's fire emissions exceed the dynamic-adjusted baseline in a Reporting Period, is compensated from the reductions Buffer Pool so the reductions account stays unbiased across the Crediting Period (see Section 5.0 and Section 9.2). A reductions debit is classified as an Avoidable or Unavoidable Reversal in accordance with the Isometric Standard. Where the Project Proponent demonstrates that the fire giving rise to the debit originated outside the Project Boundary and was not caused by the project activity or by a failure of project management, the Reversal is Unavoidable. A Project that credits reductions only contributes to and is compensated from the reductions Buffer Pool on the same basis.
• For an Avoidable Reversal, Isometric cancels Credits from the corresponding Buffer Pool equal to the loss and the Project must replenish them before further Credits are issued. Replenishment comes from Credits of the same type generated in the next Reporting Period during the Crediting Period. After the Crediting Period, replenishment comes from Credits generated by other projects operated by the Project Proponent, or from Credits of equivalent quality sourced at the Project Proponent’s expense and accepted by Isometric.
• For an Unavoidable Reversal, Isometric cancels Credits from the corresponding Buffer Pool equal to the loss.
• Where a Buffer Pool has insufficient Credits to compensate a Reversal, all further Credits issued to the Project Proponent are assigned to that Buffer Pool and cancelled until the Reversal is fully compensated, in accordance with the Isometric Standard .

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

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

Upon detection of a Reversal, Project Proponents must thoroughly investigate, initiate adaptive management to minimize losses, and implement mitigation actions to reduce future risks of Reversal.

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

A loss of credited woody-biomass stock greater than 1% of the cumulative removals credited to the Project, measured against the total Removal Credits issued, must be reported, investigated and compensated, in accordance with the Isometric Standard. On detection, the following applies:

• Within one month, the Project is notified and must investigate and confirm whether the Reversal has ended; where it has not, the Project must 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.
• Within one year, the Project must 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). 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, in accordance with the Isometric Standard. A Reversal on the scale of 20% of the cumulative removals credited to the Project requires field sampling to quantify the remaining woody-biomass stock.

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 accuracy and consistency: Every mapped product must meet a minimum accuracy and be applied identically to baseline and project.
• Appropriate safeguards: Fire management must do 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 Donor Zone transition date determined for each year (see Section 4.2.4), used to establish the historical baseline and the fire-return interval.
• The burned-area mapping method and its accuracy assessment, characterized against independent reference observations, which may be higher-resolution imagery or ground or aerial point data, to at least 80% overall accuracy, with the confusion matrix and the omission and commission rates reported, with the sample design determined on the same basis as Section 4.2.2.
• The Seasonal Cutoff Date, the Donor Zone transition-date series from which it is derived, and its calculation as the mean across the Historical Reference Period.

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.
• The characterized omission and commission rates propagated into the uncertainty estimate, with a per-period data-quality check.
• Active-fire products, for example MODIS or VIIRS hotspots, used to cross-check the post-fire burned-area maps.
• The Donor Zone transition date for each year in the Reporting Period, and any reclassification of burned area applied under Section 4.2.4.

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 over the Reporting Period and the matched control's change.
• A review of the control pixels for continued suitability, replacing any that no longer match.
• Through the removals Monitoring Period, continued monitoring of the woody-biomass stock for reversals, with no new Removal Credits issued.

Verification Requirements

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

Validation Requirements

• The delineation of the Leakage Monitoring Zone, extending 10 km from the project boundaries, 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.

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 and after applying the same dynamic adjustment used for the Project's baseline, 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).

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.

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.

The Project must reconcile its 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 donor area re-evaluated for land-cover change and updated each Reporting Period, dropping pixels that change land cover or become other carbon projects with managed fire regimes, and excluding the leakage belt.

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 organised 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. 2024⁵⁵ and apply across about 400 to 1500 mm mean annual rainfall. The northern-Australian values are the published Tier 2 defaults of the Australian Savanna Fire Management Methodology 2026⁵⁶, with emission factors from Hurst et al. 1994⁵⁷ and Meyer et al. 2012⁵⁸. Further regions are added once their parameters are validated by field derivation.

Eligible miombo must have at least 10% cover of typical miombo taxa, more than 30 trees per hectare (stems 5 cm diameter at breast height or larger), and more than 1000 shrubs taller than 1 m per hectare. Closed forest (canopy above 80%) is excluded, because fire does not carry consistently. Miombo rainfall bands are expressed as mean annual rainfall; the northern-Australian bands below are expressed as mean annual wet-season rainfall.

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).

Table A1. Vegetation Fuel Types 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 burnt⁵⁹. Pool means for coarse woody debris is 0.55, 0.52 and 0.39, and heavy woody fuel 0.91, 0.99 and 1.60, 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.

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.

Table A4. Northern Australian shrub, coarse and heavy fuel load (tonnes of dry matter per hectare), which do not vary with time since fire.