Cropland Management

This Module provides the requirements and procedures for the quantification of net carbon dioxide equivalent (CO₂e) removal from the atmosphere via improvements to soil organic carbon (SOC) stocks in cropland systems. This Module sits under the Improved Soil Management Protocol and must be read in conjunction with it.

Cropland soils represent one of the largest opportunities for terrestrial carbon sequestration¹. Globally, agricultural soils have lost an estimated 133 Pg of organic carbon since the onset of widespread cultivation, and a substantial fraction of this historical loss is recoverable through improved management practices². Recent estimates suggest that improved cropland management alone could sequester between 0.28 and 1.85 Pg CO₂ yr⁻¹ globally, placing cropland SOC enhancement among the most scalable nature-based climate solutions available³.

This Module is process-agnostic with respect to the specific management practices employed. Eligible project activities encompass any practice or combination of practices that results in a demonstrable net increase in SOC stocks against a counterfactual baseline, provided all eligibility and safeguarding requirements set out in this Module and the parent Protocol are met. This approach reflects the wide diversity of cropping systems, soil types, and climatic conditions encountered globally, and recognizes that effective SOC enhancement strategies will vary substantially by region and context. Using common terminology for SOC-enhancing practices, examples of such eligible activities include, but are not limited to:

• Cover cropping: Growing crops during fallow periods to increase organic matter inputs to the soil 
• Reduced or no-till: Minimizing soil disturbance to preserve existing SOC stocks and promote accumulation 
• Compost and organic matter application: Adding external organic inputs to increase soil carbon content 
• Biostimulant and biological inoculant application: Use of microbially-based or biochemical products that stimulate plant growth and root exudation, promoting SOC formation pathways

In addition to climate mitigation, improvements to SOC in cropland systems can provide environmental and social co-benefits, including enhanced soil water retention and drought resilience, improved crop yields and long-term agricultural productivity, reduced dependency on synthetic fertilizers, mitigation of erosion and nutrient runoff into waterways, and support for soil biodiversity ^4,5,6,7. Carbon finance presents a meaningful opportunity to overcome the adoption barriers, including implementation costs, agronomic support, and the absence of financial incentives, that have historically constrained farmer uptake of SOC-enhancing practices.

Project Proponents must meet all the requirements set out in the Improved Soil Management Protocol and relevant Modules, as well as the requirements set out in the Module. Where this Module contains requirements that duplicate or conflict with those in other Modules, this Module takes precedence.

Throughout this Module, the use of “must” indicates a requirement, whereas “should” indicates a recommendation.

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

• The Isometric Standard
• Improved Soil Management Protocol v1.0, Isometric
• 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 of this Module 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 — Lifecycle Assessment — Principles & Framework
• ISO 14044: 2006 — Environmental Management — Lifecycle Assessment — Requirements & Guidelines

Additional principles that were considered in the development of this Module include:

• The Core Carbon Principles of The Integrity Council for the Voluntary Carbon Market, v1.1, ICVCM, 2024
• Criteria for High-Quality Carbon Dioxide Removal, Carbon Direct & Microsoft, 2025

This Module was developed based on the current state of the art and publicly available science regarding cropland soil organic carbon dynamics and land management interventions. This Module aims to be scientifically stringent and robust. We recognize that some requirements may exceed the status quo in the voluntary carbon market and that there are numerous opportunities to improve the rigor of this Module as new approaches and techniques emerge.

Additionally, this Module will be reviewed when there is an update to published scientific literature, government policies, or legal requirements which would affect net CO₂e removal quantification or the monitoring guidelines outlined in this Module, or at a minimum of every 2 years.

In addition to the requirements outlined in the Improved Soil Management Protocol, Projects are subject to the following applicability requirements, which must be demonstrated in the Project Design Document.

This Module applies to land that is under active cropland management at the time of project initiation, or that has been under cropland management within the 5 years immediately prior to project initiation. In the context of this Module, cropland is defined as land used for the cultivation of annual or perennial crops, including arable land under crop rotation, fallow land within an active rotation cycle, and land under permanent crops such as orchards, vineyards, or plantations of non-timber commodity crops. Grazing land and pastoral systems are excluded from this Module.

The following land types are excluded from this Module and must not be included within the Project Boundary:

• Land that held native ecosystem cover, including native grassland, native forest, wetlands (terrestrial or tidal, including peatlands, marshes, and mangroves), or other high-conservation-value habitat, at any point within the 10 years prior to project initiation. Projects must provide evidence that land use change from native ecosystem cover did not occur within this lookback period, using remote sensing imagery, land cover classification data, or equivalent authoritative sources.
• Land classified as wetland, peatland, or organic soil (histosol), as evidenced by national or regional soil maps, land cover classification, or site-specific soil survey data. This exclusion applies regardless of whether the land is currently under cropland management.
• Land that was converted from forest, woodland, or shrubland to cropland within the 10 years prior to project initiation, unless the Project Proponent can demonstrate that the conversion was the result of a documented natural disaster or is consistent with non-industrial common practice in the region unrelated to carbon market activities or incentives. Evidence must include historical land ownership records, remote sensing data, land manager attestations, or traditional ecological knowledge documentation.
• Land under active litigation or dispute regarding ownership, tenure, or use rights, unless the dispute has been resolved prior to project validation.

• Following requirements under the Protocol, Project Proponents should implement practices which can provide additional ecological benefits in addition to increasing carbon sequestration. SOC-enhancing land management practices, including cover cropping, reduced tillage, and organic amendment application, are encouraged where they deliver co-benefits for soil health, water quality, and habitat connectivity in addition to their primary carbon benefit.
• The Project must support the livelihoods of farmers and land users enrolled in or affected by project activities. Support must be documented in the Project Design Document. See Section 6.6.1.3 of the Improved Soil Management Protocol for requirements on revenue sharing.

Additionally, projects under this Module must meet all The Project must not result in net disturbance of existing soil carbon pools through tillage or soil inversion beyond that occurring in the baseline scenario. Where cultivation is required as part of project establishment, soil inversion should be limited to 25 cm depth. This recommendation applies to areas where new management practices are being established as part of The Project and does not apply to any continuing agricultural activity that was occurring at the same depth prior to project initiation.

• The Project must provide a minimum of 40 years of SOC storage in the project area, as defined by the length of the Project Commitment Period set out in Section 5.1. The primary commitment is to the maintenance of cumulative SOC stocks at or above the level corresponding to credits previously issued. Project Proponents may adapt the specific land management practices implemented during the Project Commitment Period, provided that any such adaptation: (i) does not breach any other applicability requirement of this Module or the parent Protocol; (ii) is documented in advance in the Monitoring Report and approved at the next verification; and (iii) is accompanied by evidence, that cumulative SOC stocks are maintained or increased. Reversion to baseline land management practices, or any practice change for which the 16th-percentile cumulative SOC stock estimate falls below the level corresponding to credits previously issued, will be treated as a reversal in accordance with the Isometric Standard.

• Project Proponents must notify Isometric of any material change to the land management practices implemented within the Project Boundary as soon as practicable, and in any event no later than the next Monitoring Report. A change is material where it alters the practices identified in the Project Design Document as responsible for generating the SOC increase, or where it could reasonably be expected to affect SOC stocks within the Project Area. Failure to notify a material practice change will be treated as a non-conformance and may result in the suspension of crediting until the change has been assessed and verified.

• The Project Proponent must provide evidence that the land management practices required to maintain SOC stocks can be sustained throughout the Project Commitment Period. This includes evidence of secure land tenure or equivalent long-term land access arrangements for the project area. Failure to maintain land tenure or equivalent access may result in the cancellation of Credits, in accordance with the Isometric Standard. Where the project involves smallholder farmers or customary land users, land tenure arrangements must be consistent with the farmer and land user rights requirements of Section 6.6 of the Improved Soil Management Protocol.

• The Project must not transform the land use to the extent that it is expected to change or systematically reduce Pre-Project Productivity. This includes displacing crops grown on the Project Site historically with different crops (e.g., transitioning from a farm rotating corn and soy to planting switchgrass), introducing additional fallow periods that were not otherwise incorporated, or planting trees. The latter activities are scoped under Isometric Agroforestry and Reforestation protocols. The Project must seek to maintain or enhance agricultural productivity on project sites.

The Project Commitment Period encompasses the Crediting Period and any Ongoing Monitoring Period commitments following the end of Crediting. The Project Commitment Period must be a minimum of 40 years and no longer than 100 years. The length of the Project Commitment Period must be set at project initiation.

For grouped projects where new areas are added to the Project over time, the Project Timeline may be staggered across the project areas to reflect different initiation times of Project activities.

• Definition. The Crediting Period is the interval between project initiation (first activity on an individual site associated with the Project) and the end of the last Reporting Period. The Crediting Period is made up of successive Reporting Periods.
• Duration. The maximum duration of the Crediting Period is 100 years.
• Credit Issuance. Credit issuances occur throughout the Crediting Period. Credits are issued upon Verification of a Reporting Period. Credits issued at each verification event represent the cumulative net CO₂e removal from project initiation to the end of the current Reporting Period, less all credits previously issued under the project.
• Renewal. The Crediting Period may be renewed up to the duration of the Project Commitment Period.
• Considerations for Grouped Projects. For grouped projects composed of multiple discrete areas, the individual Crediting Periods may be staggered across individual sites to reflect the different timings of project activity initiation within the sites.

• Definition. The Reporting Period is the interval of time over which removals are assessed and crediting calculations are updated. The first Reporting Period starts at the beginning of the Crediting Period. Subsequent Reporting Periods begin at the end of the previous Reporting Period.
• Duration. The minimum duration of a Reporting Period is one year. The maximum duration of a Reporting Period is five years. Project Proponents may request an alternative length of the Reporting Period provided they submit suitable justification for the deviation (e.g., evidence of carbon stock impacts of intervention anticipated to take longer during establishment; matching relevant cultivation cycles).
• Verification. Verification of project activities by a third-party VVB is conducted for each Reporting Period (see Section 7.2 of the Improved Soil Management Protocol).
• Last Reporting Period. Project Proponents must indicate the last Reporting Period to be submitted for Verification. Failure to initiate a Verification within 5 years of the previous Reporting Period or request an extension will conclude the Crediting Period.

• Definition. The Ongoing Monitoring Period is an optional interval between the end of the Crediting Period through the end of the Project Commitment Period. An Ongoing Monitoring Period is only required if the Crediting Period is less than the length of the Project Commitment Period set at project initiation (minimum 40 years). This may be the case if an enrolled property chooses not to renew their contract to the full Project Commitment Period Length.
• Duration. If the Crediting Period is less than the Project Commitment Period, the Ongoing Monitoring Period must encompass the length of time from the end of the Crediting Period to the end of the Project Commitment Period.
• Monitoring. Monitoring for Reversals is the responsibility of the Project Proponent and must follow the same requirements for the measure re-measure approach used for quantification of CO₂e_stored (Section 9.1.2). If monitoring is not possible (e.g., do not have access to land), the area is considered to experience a full Reversal.

The Durability of Credits is set to one half the length of the Project Commitment Period. As such, the Project Commitment Period must be set at the time of PDD submission. If the initial Crediting Period is less than the Project Commitment Period, the Crediting Period may be extended up to the duration of the Project Commitment Period. If the initial Crediting Period is not extended, the remainder of the Project Commitment Period must consist of an Ongoing Monitoring Period. For grouped projects, the Project Commitment Period must be the same duration for all areas, but Crediting Period and Ongoing Monitoring Period duration may vary across individual areas to reflect differences in land tenure.

• Definition. The indefinite period of time after the Project Commitment Period has ended.
• Long-term durability. Projects must provide details of how the project design/activities will encourage long-term maintenance and sustainability of soil carbon stocks after the Project Commitment Period to prevent Reversals after the Project ends.

Project A is a smallholder cropland management project. It sets a Project Commitment Period of 40 years, with an initial Crediting Period of 20 years for all enrolled properties. At the end of the initial Crediting Period, most of the properties renew their enrollment for another 20 years and continue to accumulate Credits for the remainder of the Project Commitment Period. The remaining properties which did not renew their enrollment enter an Ongoing Monitoring Period to monitor for Reversals for the remaining 20 years of the Project Commitment Period. The Project Proponent is responsible for quantification and monitoring through all of the Project Area for the full duration of the Project Commitment Period, and the reported activities are verified by a Validation and Verification Body (VVB). All Credits from Project A have a durability of 20 years, equivalent to half of the 40 year Project Commitment Period.

Project B is a cropland project which sets a Project Commitment Period of 60 years, which fully consists of the Crediting Period. All Credits from Project B have a durability of 30 years, equivalent to half the 60 year Project Commitment Period.

Project C is a grouped cropland management project which sets a Project Commitment Period of 40 years, entirely composed of crediting periods. It is composed of two cohorts of enrolled properties - Group C1 which started in 2025 and Group C2 which started in 2026. The Project Commitment Period for Group C1 would run from 2025 to 2065, while the Project Commitment Period for Group C2 would run from 2026 to 2066. All credits would have a 20 year durability.

Project interventions involve deliberate changes to cropland management practices that may affect local ecosystems, communities, and land users beyond the direct project benefit. Project Proponents must identify and assess the environmental and social risks associated with all implemented land management practice changes in adherence with the requirements laid out within the Improved Soil Management Protocol. The following risks are specific to cropland management activities and must be addressed in the Project Design Document:

• Food security and agricultural productivity
  • Project Proponents must demonstrate that implemented practice changes do not compromise the agricultural productivity of project land or the food security of communities dependent on it.
  • Where practice changes involve reductions in fertilizer use, changes to crop rotation, or the introduction of cover crops, the Project Proponent must provide evidence that yields are not materially reduced in a manner that adversely affects local food systems or farm livelihoods.
• Agrochemical and amendment use
  • Project Proponents applying organic amendments, compost, biochar, or other soil inputs must assess the risk of soil and water contamination arising from their use.
  • Risk assessments must consider the composition of all applied amendments, including potential contaminants, and the cumulative effect of repeated applications over the project lifetime.
  • Where amendments are subject to regulations, this must be evidenced via demonstration of adherence to all relevant regulations.
• Soil health and land degradation
  • Project Proponents must demonstrate that implemented practice changes do not cause net soil degradation, including compaction, erosion, salinization, or acidification, relative to the baseline scenario.
  • Monitoring plans must include indicators of soil health beyond SOC stocks where degradation risks are identified.
• Erosion control
  • Project Proponents must assess the risk of soil erosion arising from or exacerbated by project activities, including any tillage operations, vegetation removal, or changes to ground cover associated with practice changes.
  • Where erosion risk is identified, Project Proponents must implement and document appropriate erosion control measures. These may include, but are not limited to:
    • The maintenance of permanent ground cover through cover cropping or mulching
    • The use of contour farming or terracing on sloped land
    • The establishment of vegetated buffer strips along watercourses
    • The avoidance of bare soil exposure during high-risk periods such as heavy rainfall or drought.
  • Erosion control measures must be documented in the Project Monitoring Plan.
• Farmer and land user rights
  • Where project activities are implemented on land managed by smallholder farmers, tenants, or customary land users, Project Proponents must ensure that participation is fully voluntary, that benefit-sharing arrangements are documented and equitable, and that land users retain the right to withdraw from the project without penalty.
  • Involuntary changes to land management practices are not permitted under this Protocol.
• Water use
  • Where practice changes involve irrigation or water-intensive amendments, Project Proponents must assess impacts on local water availability, including access to clean water for surrounding communities and existing water-intensive operations in the project area.

SOC-enhancing land management practices frequently require upfront investment in new inputs, equipment, and agronomic expertise, while delivering climate benefits that are not directly captured by commodity markets. Project Proponents must demonstrate that the adoption of project activities is contingent on Carbon Finance, i.e., that the practices would not be economically viable without the revenue generated from credit issuance. Where project activities generate revenue from commodity production or other non-carbon sources within the project area, Project Proponents must demonstrate that this revenue alone is insufficient to make the project financially viable, in accordance with the financial additionality requirements of the corresponding section of the Isometric Standard. Projects must not occur in regions where SOC-enhancing practices are already being driven to adoption by market demand, agricultural policy, or regulatory requirements that would lead to equivalent practice changes without Carbon Finance.

The system boundary for cropland management projects encompasses all GHG sources, sinks, and reservoirs (SSRs) associated with the implementation of SOC-enhancing land management practices on eligible croplands. The system boundary must be defined in accordance with Section 8.1 of the Improved Soil Management Protocol and the requirements below.

A cradle-to-grave GHG Statement must be prepared encompassing the GHG emissions and removals relating to all activities within the system boundary. GHG emissions and removals associated with The Project may be direct emissions from a process, or indirect emissions from combustion of fuels, electricity generation, or other sources. The Project Proponent is responsible for identifying all sources of emissions directly or indirectly related to project activities.

Any emissions from sub-processes or process changes that would not have taken place without the CDR Project must be fully considered in the system boundary. Any additional activity that ultimately leads to the issuance of Credits should be included in the system boundary.

The system boundary must include all relevant GHG SSRs controlled and related to The Project, as set out in Table 1 of the Improved Soil Management Protocol. The following module-specific considerations apply to cropland management projects:

Cropland management projects are implemented on land under active agricultural production, meaning that certain operational activities (e.g., planting, harvesting, fertilizer application, tillage passes) were occurring prior to, and may continue alongside, project activities. Activities or portions of activities that were already occurring in the baseline and would have continued to occur without the cropland management project may be omitted from the system boundary, subject to the conditions below.

For the purpose of this provision, an "activity" may refer to an operational sub-unit, such as an individual equipment pass, a single fertilizer application event, or a discrete field management step, where such sub-units can be cleanly delineated by equipment, timing, and physical scope. Where a pre-existing activity is partially modified by The Project (for example, a field operation whose frequency is altered, or an application event whose inputs are changed), the activity must be partitioned into:

• Portions that are materially unchanged by The Project, which may be excluded from the system boundary under this provision; and
• Portions that are new, extended, or altered as a result of The Project, which must remain within the system boundary and be quantified in accordance with Section 9.5 of the Improved Soil Management Protocol.

An activity, or portion of an activity, may only be excluded where the Project Proponent can demonstrate all of the following:

• It was occurring as part of routine farm operations prior to project activities;
• It will continue to be represented in the counterfactual assessment;
• It would have continued to occur in the absence of the cropland management project; and
• Its scope, frequency, equipment, inputs, and intensity are not materially altered as a result of project activities.

Evidence supporting these conditions must be provided in the PDD. This must include either:

• Historic farm records documenting the activity prior to project start. Acceptable records include management logs, operational records, equipment usage records, invoices, or equivalent documentation; or
• A signed affidavit from the relevant operator (e.g., farmer, land manager, or equivalent party) confirming the activity was part of routine operations prior to The Project.

And the following:

• A signed affidavit from the relevant operator confirming:
  • The equipment, timing, cadence, and intensity of the activity are not materially changed as a result of The Project; and
  • The activity would have continued at a comparable level absent The Project.

Where these conditions are met, only the emissions associated with the activity as it would have occurred in the baseline may be excluded. Any incremental emissions attributable to The Project must remain within the system boundary and be accounted for in the relevant emissions section.

The following SSRs are particularly relevant to cropland management projects and must be assessed:

• Cover crop seed and establishment: Where cover crops are introduced as a project activity, the embodied emissions associated with seed production, transport, and planting operations must be included.
• Changes to tillage operations: Where tillage practices are reduced or eliminated as part of The Project, the reduction in fuel use is not credited within the removals calculation (in accordance with Section 8.2.1 of the Improved Soil Management Protocol). It may be credited separately as an emission reduction under the Agricultural Practices Reductions Module, subject to the requirements set out in that Module. However, where new or additional tillage operations are introduced (e.g., for cover crop termination or soil preparation), associated fuel emissions must be included.
• Organic amendment application: Where compost, manure, biochar, or other organic amendments are applied as part of The Project, embodied emissions associated with their production, processing, and transport to site must be included.
• Biostimulant and inoculant application: Where microbial inoculants or biostimulants are introduced as a project activity, all associated production, transport, and application emissions must be included.
• Changes to fertilizer regimes: Where nitrogen fertilizer application rates or types increase as a result of The Project, the emissions associated with the new fertilizer regime must be quantified. This includes both embodied emissions of the fertilizer product and direct and indirect N₂O emissions from its application.

In accordance with the Improved Soil Management Protocol, the following are excluded from the system boundary for cropland management projects:

• Aboveground woody biomass and belowground woody biomass carbon pools (these terms must be set to zero in Equation 3 of the Improved Soil Management Protocol, as specified in Section 9 of this Module);
• Deadwood, litter, and non-woody herbaceous biomass carbon pools, as these are considered transient;

The baseline scenario for cropland management projects assumes that the SOC-enhancing management practices associated with The Project do not take place and that pre-project land management continues under business-as-usual conditions throughout the Crediting Period. The baseline must be defined in accordance with Section 8.2 of the Improved Soil Management Protocol and the requirements below.

The counterfactual represents the trajectory of SOC stocks that would have occurred in the absence of The Project, under continuation of pre-project management practices. Cropland management projects must assess the counterfactual using one of the approaches specified in Section 9.2 of this Module:

• Counterfactual assessment via measurement of control plots (for projects using a measure-remeasure quantification approach); or
• Counterfactual assessment via modeling (for projects using a measure-model quantification approach).

The counterfactual must be project-specific and reflect the land management practices, soil types, and climatic conditions of the project area. It must be dynamically updated at each Reporting Period using the most recent available data, in accordance with the requirements set out in Section 9.2.

In accordance with Section 8.2.1 of the Improved Soil Management Protocol, improved cropland management interventions may involve reductions in activities that result in emissions, such as reduced CO₂ from diesel use in tractors corresponding to fewer passes in a no-till intervention, or avoided N₂O emissions from reduced fertilizer inputs to soils.

These reductions and avoidances must not be counted separately or incorporated into the estimate of removals as part of GHG Accounting. The scope of permissible emissions reductions and rules for how to account for them are scoped in the Agricultural Practices Reductions Module. Emissions outlined in Table 1 of the Improved Soil Management Protocol and reported in accordance with Section 9.5 must be strictly positive.

This is in keeping with principles of conservativeness and maintaining consistency with Isometric's approach to GHG Accounting for removals as outlined in the GHG Accounting Module v1.1.

Leakage emissions, $CO_2e_{Leakage}$, occur when project activities lead to emissions that occur outside the system boundary of cropland management projects. For cropland management projects, the primary leakage risk is market-mediated leakage. This risk stems from project-induced reductions in agricultural production that may lead to land conversion and associated GHG emissions elsewhere to meet the supply shortfall. As a principle, Projects should seek to maintain or enhance yields from baseline levels.

Three key types of leakage can theoretically occur for cropland management projects:

• Activity-shifting leakage. Cropland management projects may displace activities in their project areas, leading to an increase in those activities outside of the project area, which may result in potential land conversion. Examples are where the local community can no longer use the whole project site for subsistence, or where part of a farmer’s productivity is displaced as a result of project activities. This type of leakage is known as “Direct” leakage as the relevant stakeholders can be identified and the activity-shifting is traceable.

• Market leakage. Cropland management projects may displace activities, which results in a reduction in supply of a commodity. Changes to the supply and demand equilibrium causes other market actors to shift their activities, leading to potential land conversion. This type of leakage is known as “Indirect” leakage because its effects cannot be isolated and measured directly. Quantifying the likelihood and potential magnitude of market leakage is complex and relies heavily on modeling and available literature.

• Ecological leakage. Project activities may lead to emissions in areas outside of the project site as a result of ecological interactions, for example unintended hydrological impacts, introduction of disease, or secondary impacts of faunal influx.

All market-leakage discounts under this Module are applied to removal credits. Where a Project concurrently generates emission reduction credits under the Agricultural Practices Reductions Module, no leakage discount is applied to those reduction credits. This allocation reflects the principle that market-mediated leakage from yield displacement is conservative to avoid the risk that emissions associated with removals are undercounted leading to over-crediting of removals.

Pre-Project Productivity ($PPP$) is defined as the annual productivity of a commodity type $c$ on Project field $f$ in relevant units (e.g., tonnes/ yr). $PPP$ must be calculated on a per-commodity basis using field-level yields indexed against regional benchmarks. This approach normalizes field-level performance against county-level (or equivalent sub-national jurisdiction) yields, controlling for weather, pest, and commodity-price variation that affects all fields in the region equally.

Field-level yield data is mandatory for the productivity assessment. The Project Proponent must obtain it through farm records, grain elevator receipts, or crop insurance records. If it can be demonstrated that none of these is available then remote sensing-based crop yield estimation may be used. Pre-Project Productivity must be based on the 5-years baseline period on field $f$ at minimum (longer baseline periods are allowed).

$PPP$ for commodity $c$ is calculated as:

$$PPP_{c,f} = \frac{1}{T_{c,f}} \sum_{t \in \{\text{baseline years with crop } c \text{ on field } f\}} \left( \bar{Y}_{regional,c} \times \frac{Y_{field,c,f,t}}{Y_{regional,c,t}} \right)$$

(Equation 1)

Where:

• $PPP_{c,f}$ is the Pre-Project Productivity for commodity $c$ on field $f$, in production units per hectare per year.
• $\bar{Y}_{regional,c}$ is the mean regional yield for commodity $c$ over the baseline period.
• $Y_{field,c,f,t}$ is the field-level yield for commodity $c$ on field $f$ in year $t$ of the baseline period.
• $Y_{regional,c,t}$ is the regional yield for commodity $c$ in year $t$ of the baseline period.
• $T_{c,f}$ is the number of years in which field $f$ was planted with commodity $c$ during the baseline period.

$PPP$ must be calculated separately for each commodity within the field's crop rotation. For example, in a corn-soy rotation with a 5-year baseline, corn $PPP$ is calculated from the years in which corn was planted and soy $PPP$ from the years in which soy was planted.

Regional data must be sourced from official agricultural statistical publications at the county level (or equivalent sub-national jurisdiction). Acceptable sources include USDA NASS county-level yield data (for the United States), FAOSTAT national yield data (where sub-national data is unavailable), or equivalent national statistical services in other jurisdictions. The data source must be documented in the PDD.

Project-Scenario Productivity ($PSP$) for commodity $c$ on field $f$ in Reporting Period $RP$ is calculated as:

$$PSP_{c,f,RP} = \bar{Y}_{regional,c} \times (1 + GR_c)^{n} \times \frac{Y_{field,c,f,RP}}{Y_{regional,c,RP}}$$

(Equation 2)

Where:

• $PSP_{c,f,RP}$ is the secular-trend-adjusted Project-Scenario Productivity for commodity $c$ on field $f$ in Reporting Period $RP$, in production units per hectare per year.
• $\bar{Y}_{regional,c}$ is the mean regional yield for commodity $c$ over the baseline period, in production units per hectare, as defined in Equation 1.
• $Y_{field,c,f,RP}$ is the field-level yield for commodity $c$ on field $f$ in the Reporting Period, in production units per hectare.
• $Y_{regional,c,RP}$ is the regional yield for commodity $c$ in the Reporting Period, in production units per hectare, drawn from the same regional yield source used for $\bar{Y}_{regional,c}$.
• $GR_c$ is the annual growth rate in productivity of the commodity type, see section 9.2.2.
• $n$ is the number of years from the midpoint of the baseline period to the midpoint of the current Reporting Period.

Where the Reporting Period spans multiple years, $Y_{field,c,f,RP}$ and $Y_{regional,c,RP}$ must each be calculated as the simple average of the per-year values across the years in which commodity $c$ was planted on field $f$ during the Reporting Period.

The productivity shortfall for each commodity $c$ in Reporting Period $RP$ is:

$$\Delta P_{c,f,RP} = \max(PPP_{c,f} - PSP_{c,f,RP}, \ 0)$$

(Equation 3)

No leakage assessment is required for commodities where $\Delta P_{c,f,RP} = 0$ (i.e., project-scenario productivity meets or exceeds the baseline).

The productivity assessment is performed independently for each Reporting Period against the original pre-project baseline. Each Reporting Period's leakage is calculated without reference to the leakage assessed in any prior Reporting Period.

Where $\Delta P_{c,RP} / PPP_c \leq 0.03$ (i.e., the productivity decline is 3 percentage points or less relative to the baseline for commodity $c$), the decline is within the de minimis materiality threshold and no leakage discount applies for that commodity in that Reporting Period. The de minimis threshold does not serve as basis from which the leakage is then assessed; if the productivity decline is greater than 3%, leakage must be applied to the total amount.

Where $\Delta P_{c,RP} / PPP_c > 0.15$ (i.e., the productivity decline exceeds 15 percentage points for any single commodity), The Project is ineligible for crediting for that Reporting Period unless a temporary exemption is granted under the exceptional circumstances provision (see below).

Where some fields within the project experience yield increases as a result of SOC-enhancing interventions, the additional production may be used to offset productivity shortfalls on other fields within the same project. This within-project netting reflects the economic reality that additional supply from fields with increased yields fields relieves the same market pressure that fields from decreased yields create.

A field may generate leakage mitigation (a productivity surplus that can offset negative leakage elsewhere in the project) only where all of the following conditions are met:

• The field is enrolled in the project and implementing SOC-enhancing interventions under the Improved Soil Management Protocol and this Module. This requirement is to provide confidence that the yield increases are a result of The Project intervention(s), not incidental.
• The yield increase exceeds 3% above the growth-adjusted Pre-Project Productivity for the commodity. This mirrors the de minimis gate on the negative side and ensures that only meaningful, above-trend increases are counted — not normal year-to-year variability or background yield growth captured by the secular trend adjustment.

The productivity surplus for commodity $c$ on field $f$ in Reporting Period $RP$ is:

$$\Delta P^{+}_{c,f,RP} = \max(PSP_{c,f,RP} - PPP_{c,f} \times 1.03, \ 0)$$

(Equation 4)

Where:

• $\Delta P^{+}_{c,f,RP}$ is the productivity surplus for commodity $c$ on field $f$, in production units per hectare. Only the portion exceeding the 3% gate is counted.
• $PSP_{c,f,RP}$ is the Project-Scenario Productivity for commodity $c$ on field $f$ (Equation 2).
• $PPP_{c,f}$ is the Pre-Project Productivity for commodity $c$ on field $f$ (Equation 1).
• The factor of 1.03 applies the 3% positive-leakage gate.

The total project-level productivity surplus for commodity $c$ is:

$$Total\ Surplus_{c,RP} = \sum_f \Delta P^{+}_{c,f,RP} \times A_{c,f,RP}$$

(Equation 5)

Where the sum is taken by scaling across the corresponding areas of all individual fields $f$ within the project ($A_{c,f,RP}$) that are eligible for positive leakage and are planted with commodity $c$ in the Reporting Period.

Positive leakage may only offset negative leakage within the same commodity. A corn yield surplus cannot offset a soy yield shortfall as these are different markets with different supply chains and different land conversion intensities. Netting across commodities would conflate distinct market responses and risk underestimating actual leakage.

Positive leakage may only offset negative leakage within the same Reporting Period. Positive leakage cannot be carried over to offset negative leakage in a subsequent (or historical) Reporting Period.

Positive leakage is applied by reducing the project-level Net Project Productivity (Equation 5) for each commodity before the IS/NL calculation is performed. This ensures that the market-response model operates on the net supply impact after within-project mitigation, rather than modeling positive and negative leakage separately and attempting to net the CO₂e outputs.

Where the productivity decline for commodity $c$ exceeds the 3% de minimis threshold ($\Delta P_{c,RP} / PPP_c > 0.03$), the Net Project Productivity is calculated on the full shortfall:

$$NPP_{c,RP} = max( Σ_f ( ΔP_{c,f,RP} × A_{c,f,RP} ) − Total\ Surplus_{c,RP} , 0 )$$

(Equation 6)

Where:

• $\Delta P_{c,f,RP}$ is the productivity shortfall for commodity $c$ on field $f$ (Equation 3), in production units per hectare per year.
• $A_{c,f,RP}$ is the area planted with commodity c on field f in the Reporting Period, in hectares.
• $Total/ Surplus_{c,RP}$ is the within-project positive-leakage surplus for commodity $c$ (Equation 5), in production units per year. The sum is taken over all fields $f$ within the project planted with commodity $c$ in the Reporting Period.

The max ensures NPP is non-negative even when within-project positive leakage exceeds the gross shortfall.

Adjusted Net Project Productivity, $aNPP$, represents the remaining productivity deficit after accounting for cropland system productivity and any external mitigation, adjusted for growth trends in commodity productivity. This value reflects the net shortfall that may trigger market responses leading to land conversion elsewhere.

$aNPP$, must be calculated using the following equation:

$$aNPP_{c} = NPP_{c} \times (1 + GR_{c})$$

(Equation 7)

Where:

• $aNPP$ is the adjusted Net Project Productivity, in appropriate units (e.g., tonnes per year).
• $NPP$ is the Net Project Productivity, representing the average annual production deficit after accounting for Project-Scenario Productivity in appropriate units (e.g., tonnes per year).
• $GR$ is the annual growth rate in productivity of the commodity type.

The annual growth rate in productivity of the commodity type and region must be assigned as part of Equation 7. This is a requirement to ensure that any likely future increases in productivity are accounted for as part of the assessment.

Growth rate must be calculated based on the following hierarchy:

1. Regional commodity specific data should be used to source yield values to determine growth rates, where available. For example from the United States Department for Agriculture for the commodity type and state;
2. National commodity specific data for agriculture from FAOSTAT may be used to source yield values to determine growth rates.

Growth rate must be calculated using the following equation:

$$GR_{c} = \left( \frac{yield_{c,t}}{yield_{c,t-x}} \right)^{1/x} - 1$$

(Equation 8)

Where:

• $GR$ is the growth rate, as a percentage.
• $yield$ is the yield as sourced from regional or national datasets, in production unit per hectare per year.
• $c$ is the commodity type
• $t$ is the most recent year of recorded yield data.
• $t−x$ is the historic year yield data.
• $x$ is the number of years between $t$ and $t-x$

Average growth rate is determined by taking the difference between yield in the most recent year of recorded data ($t$) and a historic year ($t−x$). Where possible $t−x$ should represent 25 years prior to $t$. Where this is not possible, a minimum of 10 years prior to $t$ is allowable.

If a recent negative shock leads to a negative growth estimate of yield growth, a value of zero should be used.

Project Proponents are required to estimate the amount of new land brought into production, $ha_{LC}$. This estimate must be informed by:

• Pre-Project Productivity of a commodity type at the project site (after accounting for system productivity);
• The estimated proportion of this productivity that would be replaced with new production via an increase in supply of the commodity type;
• The increase in supply that would result in new land being brought into production; and
• The yield of new land being brought into production.

The new land brought into production must be calculated separately for each commodity type being displaced as a result of the Project.

For each commodity $c$ with $aNPP_{c,RP} > 0$, the hectares of induced land conversion are calculated as:

$$ha_{LC,c,RP} = \frac{aNPP_{c,RP} \times IS_c \times NL_c}{Y_{NL,c}}$$

(Equation 9)

Where:

• $ha_{LC,c,RP}$ is the induced land conversion for commodity $c$ in the Reporting Period, in hectares.
• $aNPP_{c,RP}$ is the adjusted Net Project Productivity, in appropriate units (e.g., tonnes per year).
• $IS_c$ is the Increased Supply fraction for commodity $c$ — the proportion of the foregone deficit that will be replaced by increased supply elsewhere, calculated from supply and demand elasticities in accordance with Section 8.4.2.1 of this Module and Appendix E.
• $NL_c$ is the proportion of increased supply for commodity $c$ that will result in new land being brought into production, sourced in accordance with Section 8.4.2.2 of this Module and Appendix E.
• $Y_{NL,c}$ is the yield on new land brought into production for commodity $c$, in production units per hectare per year, determined in accordance with Section 8.4.2.3 of this Module.
• $c$ is commodity type.

Where the Project falls into regions for which Isometric has provided default IS and NL values in Appendix E, those default values must be used. For all other regions, values must be sourced from literature following the procedures set out in Appendix E.

Increased Supply ($IS$) is the proportion of foregone productivity that will be replaced by increased supply elsewhere. This is underpinned by the premise that foregone production will not necessarily be replaced in totality by increased supply elsewhere as a result of elasticities of supply and demand. Global markets for commodities have been assumed for the purposes of the leakage assessment.

Estimates for IS are determined using the following equation:

$$IS = \frac{\varepsilon_{s,c}}{\varepsilon_{s,c} + \lvert \varepsilon_{d,c} \rvert}$$

(Equation 10)

Where:

• $IS$ is increased supply, as a percentage.
• $\epsilon_{s}$ is elasticity in supply, as a ratio.
• $\epsilon_{d}$ is elasticity in demand, as a ratio.
• $c$ is commodity type.

Isometric has carried out a literature review of $\epsilon_{s}$ and $\epsilon_{d}$ values for certain regions. Values for $\epsilon_{s}$ and $\epsilon_{d}$ for these regions are provided in Appendix E. Where The Project falls into these regions, the default values provided must be used. This is because understanding which values to use from literature is challenging as academic papers are typically not written with this purpose or audience in mind. Isometric has completed this work for certain regions to lessen this complexity and provide consistency across projects.

The default values also serve as an example of appropriate values to select from the literature for other regions; however, it should be noted that the quality of research differs across regions. For all other regions, values for $\epsilon_{s}$ and $\epsilon_{d}$ must be sourced from literature. The procedure and requirements for sourcing default values for $\epsilon_{s}$ and $\epsilon_{d}$ are set out in Appendix E.

$NL$ considers the percentage of increased supply that will result in new land brought into production for the commodity type. This is underpinned by the premise that not all increased supply will result in new lands being brought into production. Some increased supply may be made up of intensification of activities and increased yields on existing production lands.

Isometric have carried out a literature review of $NL$ values for certain regions. Values for $NL$ for these regions are provided in Appendix E. Where The Project falls into these regions, the default values provided must be used. The procedure and requirements for sourcing default values for NL are set out in Appendix E.

The default values also serve as an example of appropriate values to select, however it should be noted that the quality of research differs across regions.

$Y_{NL,c}$ considers the yield on new land brought into production for commodity $c$. This is assessed based on the observed productivity in the region in the pre-project period. Here, the value of the regional mean yield ($\bar{Y}_{regional,c}$) used in Section 8.3.2.1 for assessing pre-project productivity must be assumed as the the value of $Y_{NL,c}$.

$EF_{Carbon \ Stock}$ must be derived from the IPCC average national aboveground biomass content of the land cover for the relevant ecosystem. Mean carbon stocks should be derived from aboveground biomass estimates in Table 3A.1.4 of the IPCC Good Practice Guidance for Land Use, Land Use Change and Forestry⁸.

Carbon stocks should be determined using the ratio of mass of CO₂ to mass of C, and carbon fraction, $CF$, specified for the ecoregion/vegetation type by the IPCC.

Total leakage emissions for the Reporting Period are the sum of leakage across all commodities with a productivity shortfall:

$$CO_2e_{Leakage,RP} = \sum_c \left( ha_{LC,c,RP} \times EF_{Carbon \ Stock} \right)$$

(Equation 11)

Where:

• $CO_2e_{Leakage,RP}$ is the total leakage emission for the Reporting Period, in tonnes CO₂e.
• The sum is taken over all commodities $c$ for which $NPP_{c,RP} > 0$.

$CO_2e_{Leakage,RP}$ is included as part of $CO_2e_{Emissions,RP}$ as set out in the Improved Soil Management Protocol. The leakage discount is applied exclusively to removal credits; no leakage discount is applied to emission reduction credits issued under the Agricultural Practices Reductions Module.

$CO_2e_{Leakage}$ is quantified for every Reporting Period.

The Project Proponent must document the baseline crop rotation for each enrolled field, including the sequence and frequency of each commodity planted over the 5-year baseline period. For each commodity in the baseline rotation, the number of plantings over the baseline period must be documented.

The Project Proponent must maintain the same set of commodities as the baseline rotation within a ±1 year tolerance over each 5-year window. For example, a baseline rotation of corn-corn-soy-corn-soy permits any combination that includes 2–4 years of corn and 1–3 years of soy within the next 5 years.

For fallow rotations, The Project can maintain the use of fallow rotations at the same historical rate. If the number of fallow rotations within the project exceeds the historical rate (e.g., a third fallow rotation within a 5-year period when the baseline featured two), the forgone production for the additional fallow rotation is assumed to equal the average production of the most productive commodity grown on the field during the baseline period.

Crop changes beyond the ±1 year tolerance are permitted if pre-registered at least 4 months before planting with supporting evidence that the change is driven by regional market trends, agronomic factors, or farm-level economic conditions unrelated to the project. Pre-registered crop changes that meet these criteria do not incur a leakage penalty.

Any unregistered crop change that results in a shift to a lower-value commodity or an increase in fallow rotations beyond the baseline rate must be treated as a productivity shortfall for leakage purposes. The forgone production is assumed to equal the average production of the most productive commodity grown on the field during the baseline period.

If a crop failure occurs in a Reporting Period, the Project Proponent must provide evidence that the failure was caused by factors outside the Project's control (e.g., extreme weather, pest outbreak). If accepted as a genuine crop failure:

• The field's yield for that Reporting Period is set to the average yield of other rotations of the same commodity during the baseline period for purposes of the productivity assessment. This prevents a crop failure from inflating the productivity shortfall.
• If no other baseline observations exist for the failed commodity, the field's yield is set to zero and the full productivity shortfall is assessed.

Where $PSP$ is significantly lower than projected due to exogenous natural causes beyond the Project's reasonable control (extreme weather, drought, flooding, region-wide pest or disease outbreaks), the Project Proponent may request that $PSP$ be evaluated in a broader regional context.

The regional indexing approach (Equations 1–2) already controls for most region-wide shocks — if the field and the region both experience a drought, the yield ratio ($Y_{field}/Y_{regional}$) is largely unaffected. The exceptional circumstances provision therefore applies to localised events that affect the project field but not the broader region.

Evidence must demonstrate that:

1. The cause of the $PSP$ shortfall is unrelated to project design, management decisions, or land-use change; and
2. Comparable agricultural systems in the surrounding region did not experience similar yield impacts during the same period.

Evidence may include regional or county-level yield statistics, meteorological or disaster records, government or insurance reports, and peer-reviewed or authoritative third-party data.

Where Isometric determines that the exogenous cause contributed to $PSP$ shortfalls, Isometric may adjust the $PSP$ by applying a counterfactual estimate that reflects the localized nature of the impact for the purposes of leakage assessment.

Setup: A project with 100 hectares enrolled in a corn-soy rotation. Baseline $PPP$ (regionally indexed): corn = 180 bu/ha, soy = 50 bu/ha. $IS$ = 0.70 (global calories, from Appendix E), $NL$ = 0.28 (US cropland, from Appendix E). $Y_{NL}$ = 150 bu/ha (corn) and 45 bu/ha (soy). $EF_{Carbon Stock}$ = 200 t CO₂e/ha.

Reporting Period 1 (Corn Year):

• PSP (regionally indexed, growth-adjusted) = 170 bu/ha. Decline = (180 − 170) / 180 = 5.6%.
• Exceeds 3% de minimis gate. Leakage assessed on the full shortfall: $\Delta P$ = 180 − 170 = 10 bu/ha.
• $NPP$ = 10 × 100 ha = 1,000 bu/yr.
• $ha_{LC}$ = (1,000 × 0.70 × 0.28) / 150 = 1.307 ha.
• $CO_2e_{Leakage,1}$ = 1.307 × 200 = 261.3 t CO₂e.

Reporting Period 2 (Soy Year):

• PSP (regionally indexed, growth-adjusted) = 46 bu/ha. Decline = (50 − 46) / 50 = 8%.
• Exceeds 3% de minimis gate. Leakage assessed on the full shortfall: $\Delta P$ = 50 − 46 = 4 bu/ha.
• $NPP$ = 4 × 100 ha = 400 bu/yr.
• $ha_{LC}$ = (400 × 0.70 × 0.28) / 45 = 1.742 ha.
• $CO_2e_{Leakage,2}$ = 1.742 × 200 = 348.4 t CO₂e.

Note: the soy-year leakage is larger than the corn-year leakage despite a smaller absolute shortfall in bushels (4 vs 10), because soy has a lower yield on new land ($Y_{NL}$ = 45 vs 150), meaning more land must be converted per unit of displaced soy production. Each Reporting Period is assessed independently against the baseline.

Reporting Period 3 (Corn Year):

• PSP (regionally indexed, growth-adjusted) = 176 bu/ha. Decline = (180 − 176) / 180 = 2.2%.
• Below 3% de minimis gate. No leakage assessed.
• $CO_2e_{Leakage,3}$ = 0.

Note: the 2.2% corn yield decline in RP3 is within the 3% de minimis threshold. The decline is treated as natural year-to-year variability rather than a project-induced productivity loss. Because the de minimis threshold is a gate (not a deduction), no portion of the 2.2% decline is subject to leakage — the entire Reporting Period is leakage-free for corn.

The net removals are calculated following the requirements within the Improved Soil Management Protocol. Under this Module, the carbon storage ($CO_2e_{stored}$) is considered to consist of soil organic carbon. Terms for aboveground and belowground woody biomass must be set to 0.

The following calculations apply equally to both Quantification Approach 1 (direct measurement and remeasurement, Section 9.1.2) and Quantification Approach 2 (biogeochemical modeling with remeasurement, Section 9.1.3). They consume the stratum level SOC stock outputs from either approach and produce the project-level total stock and cumulative stock change used for credit issuance.

The project-level total carbon stock at any time $t$ is expressed as total carbon mass of soil organic carbon across all strata within the quantification unit:

$$CO_2e_{stored,t} =
        \sum_{j=1}^{J}
        \left(
            SOC_{j,t} \times A_j
        \right)\times \frac{44}{12}$$

(Equation 12)

Where:

• $CO_2e_{stored,t}$ is the project-level carbon stock at time $t$ in tonnes CO₂e
• $SOC_{j,t}$ is the SOC stock in stratum $j$ at time $t$ in tonnes C ha^-1
• $A_j$ is the area of stratum $j$ in ha
• $J$ is the total number of strata
• $\frac{44}{12}$ is the mass ratio for converting carbon to CO₂e

$SOC_{j,t_n}$ is substituted as follows depending on the quantification approach and crediting event type. All substitution targets are calculated on an equivalent soil mineral mass (ESM) basis using a fixed-increment formulation across the depth increments required under Section 9.1.2.4, with parameters estimated according to the differential sampling intensity by depth set out in that Section:

• Approach 1, all events: $SOC_{j,t_n}$ from Equation 16 (ESM-corrected direct measurement on a mineral-mass basis).
• Approach 2, interim crediting events: $SOC^{model}_{j,t_n}$ from Equation 17 (mineral-mass model output using baseline reference mineral mass; 1.2 SD uncertainty discount applies in accordance with Section 9.1.3.1).
• Approach 2, true-up events: $SOC^{model,ESM}_{j,t_n}$ from Equation 18 (ESM-corrected mineral-mass model output using true-up-updated reference mineral mass; 1.0 SD uncertainty discount applies in accordance with Section 9.1.3.1).

The differential sampling intensity by depth means that the variance contributions to $SOC_{j,t}$ for the 0–30 cm and 30–100 cm components of the depth profile are characterized separately and must be propagated through the Section 9.1.1 Monte Carlo simulation independently before being combined at the stratum level. The aggregation across strata in Equation 12 itself is unaffected.

The values of $CO_2e_{stored,t}$ are then used in Equation 2 of the Improved Soil Management Protocol to derive the Reporting-Period change in stored carbon ($CO_2e_{stored,RP}$), which feeds Equation 1 of the Protocol to calculate net removals ($CO_2e_{Removal,RP}$).

Uncertainty in the net CO₂e removal estimate arises from multiple sources under both quantification approaches. All uncertainty must be propagated through the net CO₂e removal calculation using Monte Carlo simulation, in accordance with the requirements below and the corresponding section of the Isometric Standard.

Monte Carlo simulation is required under this Protocol for the following reasons:

• It does not require the assumption that measurement or model errors are uncorrelated beyond the field level, making it more appropriate for the structured spatial and temporal error patterns present in both direct measurement campaigns and biogeochemical SOC models;
• It can directly estimate the correlation of errors across years, avoiding underestimation of uncertainty on shorter time horizons where annual errors may be positively correlated;
• Where implemented as a hierarchical bootstrap (see below), it requires minimal assumptions about the shape of the underlying data distributions, which is particularly important for paired SOC stock differences that are typically non-normal and exhibit short-range spatial autocorrelation.

Monte Carlo simulation must be implemented as follows:

• An appropriate number of iterations must be performed. Project Proponents must demonstrate convergence by showing that the uncertainty estimate does not materially change (> 1%) with additional iterations.
• Each iteration of the Monte Carlo simulation must be independent of other iterations. Within a given iteration, the sources of uncertainty contributing to the credit estimate must be sampled jointly, with correlations between sources preserved in accordance with the correlation requirements set out below; the requirement of inter-iteration independence does not imply that sources sampled within an iteration are independent of one another. The materially uncertain sources sampled at each iteration may include
  • (i) input feature uncertainty, sampled from the probability distributions of materially uncertain inputs;
  • (ii) model parameter or weight uncertainty, sampled from the posterior distribution for Bayesian models or via bootstrap resampling for frequentist methods;
  • (iii) ensemble-member sampling, where the ensemble represents diversity in training data, hyperparameters, or architecture;
  • or (iv) any combination of these, depending on which sources are material to the model in use. The relevant sources differ by approach as specified below.
• Input probability distributions must be justified with reference to empirical data or published literature and must not be assumed normal without justification. Where distributions are non-normal, appropriate alternative distributions must be used.
• Correlations between uncertain inputs must be explicitly assessed and, where material, accounted for through appropriate joint sampling procedures rather than assuming independence.
• The correlation structure of errors across time periods must be characterized and incorporated into the simulation, to avoid underestimation of cumulative uncertainty over multi-year Reporting Periods.

The following uncertain inputs must be characterized and included in the Monte Carlo simulation:

• Both approaches: field sampling variance (derived from the observed standard deviation of within-location stock changes across paired measurements within each stratum), with appropriate adjustment for spatial autocorrelation among sampling locations such that the effective sample size used in standard-error calculations reflects the empirical SOC residual variogram (see §9.1.1 Hierarchical Bootstrap, Step 3); laboratory analytical uncertainty characterized in accordance with §9.1.2.7 ("Quantification of analytical uncertainty for Monte Carlo propagation"); bulk density measurement uncertainty characterized separately and propagated jointly with SOC concentration uncertainty; and baseline measurement uncertainty at $t_0$, which propagates through all subsequent cumulative delta calculations.

  • In the event that any sampling locations are substituted during a given Reporting Period (see Section 9.1.2.2.4), the unpaired component of variance for these locations must be propagated separately from the paired component, with the two components using sample-count weighting. The procedure for this must be clearly documented.
  • Field sampling variance must be characterized separately for depth increments sampled across all locations (0–15 cm and 15–30 cm) and depth increments sampled only at deep-core locations (30–60 cm and 60–100 cm), reflecting the differential sample size by depth. The two variance components must be propagated independently within the Monte Carlo simulation and combined at the stratum-level SOC stock estimate.
  • Survey uncertainty arising from estimating stratum-level mean SOC stocks from a finite sample of locations or modeled units within each stratum must also be characterized and propagated, distinct from per-location measurement and analytical uncertainty.

• Model-remeasure only: interim crediting events: model prediction error (characterized by RMSE from the validation dataset); parameterization uncertainty; and temporal error correlation across modeled time steps between resampling campaigns. Note that at interim events there is no measured bulk density, so the reference bulk density $BD_{j,ref}$ carries uncertainty from the original baseline measurement.

• Model-remeasure only: true-up events: model prediction error and parameterization uncertainty as updated through the true-up procedure in Section 9.1.3.2.4. Measured bulk density from the resampling campaign replaces the reference value, reducing uncertainty relative to interim events.

The conservative estimate used for credit issuance must correspond to the 16th percentile (one standard deviation below the mean) of the distribution of net CO₂e removal estimates across all simulation iterations, in line with the Isometric standard. Where the 16th percentile estimate is negative, no credits may be issued for that Reporting Period.

Where model-based error propagation is used under Measure-Model, the coverage of model predictions must be evaluated at each true-up event. Where coverage is materially below the nominal confidence level, the input uncertainty distributions must be revised before the simulation is rerun. This requirement does not apply under Measure-Remeasure.

The uncertainty information reported at each verification must include the items specified in Section 7.6 of the Improved Soil Management Protocol, and must document which input distributions were used, the correlation structure assumed, and evidence of convergence.

For propagating sampling and analytical uncertainty under both Approach 1 and Approach 2, Project Proponents are recommended to implement a hierarchical bootstrap Monte Carlo procedure, since it makes minimal assumptions about the shape of the underlying data distributions and naturally captures within-stratum and between-stratum variability. The procedure has three steps:

Step 1 — Location resampling within strata. Within each stratum $j$, resample sampling locations with replacement, drawing $n_j$ locations where $n_j$ is the number of sampling locations in stratum $j$. Resampling is performed on locations (not on individual cores or depth increments), so that the paired structure ($t_0$ and $t_n$ measurements at the same location) is preserved within each bootstrap iteration. Location resampling at stratum $j$ may only be applied where rarefaction analysis (see Step 4 below) demonstrates that $n_j$ supports stable estimates; otherwise the project must fall back to a parametric Monte Carlo at the stratum level with explicit assumptions documented in the Monitoring Report.

Step 2 — Observation perturbation. For each resampled location, perturb the measured SOC concentration and soil mineral mass per depth increment by drawing from the analytical uncertainty distribution characterized under §9.1.2.7 (see Fix 5 below). Where a location is drawn more than once in a single bootstrap iteration, the same analytical perturbation must be applied to all repeated draws of that location, so that analytical noise is not double-counted as sampling variance. Where significant spatial autocorrelation is detected (see Step 3), the effective sample size $n_{eff}$ replaces $n_j$ in the resampling count for that stratum.

Step 3 — Spatial autocorrelation test. Within each stratum, test for spatial autocorrelation in the location-level paired SOC stock differences using Moran's $I$ or the empirical SOC residual variogram. Where significant autocorrelation is detected at the typical inter-location distance, compute an effective sample size $n_{eff}$ for that stratum using a documented adjustment (e.g., Cressie's variance-inflation formula based on the fitted variogram), and substitute $n_{eff}$ for $n_j$ in Step 1.

Step 4 — Stratum and project aggregation. Compute the stratum-level mean SOC stock change from the resampled locations. Compute the project-level mean as the area-weighted average of stratum-level means. Repeat steps 1–4 across iterations until the 16th-percentile estimate (see below) has converged within the 1% threshold of §9.1.1.

Parametric fallback. Where the number of sampling locations in a stratum is too small to support stable bootstrap estimates, or where the rarefaction analysis fails (see below), Project Proponents may fall back to a parametric Monte Carlo simulation under §9.1.1, drawing from a stratum-level distribution justified against empirical data or literature. The fallback must be flagged in the Monitoring Report and a conservative additional uncertainty penalty applied at Isometric's direction.

Direct measurement is used to quantify changes in SOC stocks under this approach. This approach is applicable where predictive models are unavailable, have not been validated for the relevant soil type or land management context, or have not been sufficiently parameterized for the project area. Project Proponents may also elect to use direct measurement where they prefer not to rely on modeled outputs for SOC stock change quantification.

Under this approach, baseline SOC stocks within the Project Area are established through direct field measurement at project initiation ($t_0$) and re-measured at each subsequent Reporting Period. The counterfactual SOC trajectory is established through paired measurement of business-as-usual control plots, as set out in Section 9.2.1. This approach is applicable to the quantification of SOC stock changes and is not used for other GHG sources or sinks within the system boundary, which are quantified separately under Section 9.5 of the Improved Soil Management Protocol.

Project Proponents must apply standard QA/QC procedures for soil inventory, covering all stages of field data collection and data management. QA/QC procedures must be documented in the Project Monitoring Plan and applied consistently across all Reporting Periods.

Project Proponents are encouraged to adopt or adapt QA/QC procedures from established published frameworks, including those produced by the Food and Agriculture Organization of the United Nations (FAO) and available via the FAO Soils Portal, the ISO soil sampling standards (including ISO~18400-104: Soil Quality --- Sampling --- Part 104: Strategies), or the IPCC Good Practice Guidance for Land Use, Land-Use Change and Forestry (2003).

For all directly sampled parameters, the Project Monitoring Plan must:

• Clearly delineate the spatial extent of the sample population;
• Specify sampling intensities, quantification unit selection criteria, and sampling stages where a multi-stage design is applied;
• Identify unbiased estimators of population parameters for use in all calculations; and
• Include a statistical analysis plan, submitted as part of the sampling plan at project validation.

The following requirements apply to all sampling and re-sampling campaigns:

• Sample locations must be georeferenced to a horizontal accuracy of 3 m (or better) to enable re-sampling at the sampling locations set at project initiation during subsequent Reporting Periods;
  • At each re-sampling, samples must be collected within a 5 m radius of the initial location set at project initiation, making sure to avoid the exact location of any prior sampling disturbance, with the coordinates of the new sample collected and recorded in the Monitoring Report;
• Intra-annual variability must be considered in the sampling design, and all sampling and re-sampling campaigns must be conducted during the same crop phase (e.g., pre-planting, post-harvest) and within +/- 30 days across Reporting Periods to ensure comparability;
• Where organic amendments have been applied, Project Proponents must delay sampling or re-sampling to the latest practicable point after the previous application and the earliest practicable point before the next application, in order to minimize confounding effects on measured SOC stocks.

Sampling must be designed to produce an unbiased, statistically defensible estimate of SOC stocks and stock changes at the project level, with a transparent and reproducible link between sampling locations and the population they represent. The choice of design is determined by the Project Proponent and justified at validation against the project's quantification approach, the heterogeneity of the project area, and the expected precision of the project-level estimate. The protocol specifies a default design but does not mandate sub-field stratification or any particular set of stratification factors where a simpler design can be justified to deliver equivalent or better project-level precision.

The Project Proponent must define a hierarchy of quantification units in the Project Monitoring Plan. At minimum, this must specify:

• The primary quantification unit — the unit at which credits are issued and uncertainty is reported. This is typically the project area or, for grouped projects, an enrolled property or farm.
• The sampling unit — the unit at which sampling locations are selected and at which a single SOC stock estimate is generated. Sampling units must nest within primary quantification units.
• Where applicable, intermediate units (e.g., field, management zone) used to organize the design.

The choice of units must reflect the management, soil, and climatic heterogeneity of the project area and the practical constraints of sampling, including for projects involving smallholders. Sub-field stratification is not required where a coarser unit can be shown to deliver equivalent or better project-level precision.

The default sampling design is stratified random sampling, in which each sampling unit is divided into strata more homogeneous in expected SOC than the unit as a whole. Project Proponents may instead propose:

• Conditioned Latin hypercube sampling (cLHS) or other covariate-informed probability sampling designs, where ancillary variables (e.g., elevation, slope, remote-sensed indices, digital soil maps) are available and shown to correlate with SOC stocks;
• Model-based sampling designs (e.g., spatial coverage sampling, geostatistical designs) where the project intends to use a model-based estimator at the project level.

Any design must be documented in the Project Monitoring Plan and justified with reference to peer-reviewed literature. Grid sampling and unstratified simple random sampling are not permitted. Project Proponents should consider, as part of the design choice, the robustness of the design to anticipated point-level data loss arising from withdrawal, sampling-access loss, or laboratory error. A design with a large number of fine-grained strata can be more sensitive to such losses than a coarser design

Strata, where used, must be delineated using the best available data on factors that influence SOC stock distribution and the response of SOC to project activities. The factors set out in this paragraph are illustrative of those typically relevant at field (10–100 ha) and landscape (100–1,000 ha) scales — climate, topography, historical land use and vegetation, parent material, soil texture, soil type, and where available, remote-sensed indicators such as bare-soil reflectance composites or vegetation indices — but are not mandatory. Project Proponents must select stratification factors based on their relevance to the specific project's heterogeneity and quantification approach, recognizing that adding additional factors yields diminishing precision returns and increases the risk that strata fall below the three-samples-per-stratum minimum. Soil maps and databases including the FAO Soils Portal, SoilGrids, or locally available digital soil maps may be used to inform stratum delineation. Field boundaries should be considered where management history aligns with them.

Where stratified random sampling is used in a paired-difference design (i.e., the same locations are resampled across Reporting Periods), stratification factors should be selected for their expected influence on the rate of SOC change under project management, not on absolute baseline SOC levels. Within-location baseline absolute SOC largely cancels out in the paired difference; residual within-stratum variance is dominated by heterogeneity in management practice intensity, climate, soil texture, and initial SOC saturation. Where the project area is homogeneous in these factors, a coarser sampling unit may deliver equivalent project-level precision to a finely stratified design.

Project Proponents must use the sample design set at project initiation for sampling during subsequent Reporting Periods. Sampling locations may be substituted over the course of the Project Commitment Period only when there are documented barriers, including:

• Loss of contractual access (e.g., disenrollment of a property)
• Active hazards that prevent access to sampling site (e.g., natural disaster)

In such a scenario, the sampling location must be replaced with another location in the same stratum selected using the same probability sampling rule that was applied at project initiation. The substitution must be clearly documented in the Monitoring Report for review as part of the subsequent Verification. Uncertainty associated with the substitution also must be accounted for following the requirements in Section 9.1.1. This substitute sampling point must then be used for the remainder of the Reporting Period.

Project Proponents must track all substitutions over the entire Project Commitment Period. If the number of substitutions within a single strata exceeds 15% of sampling points within that strata within a single Reporting Period, or cumulatively exceeds 25% of the sampling points within that strata over the Project Commitment Period, the strata must be re-evaluated in accordance with Section 9.1.2.2.5. This must include a revised power analysis following Section 9.1.2.3.

If the sampling location is only temporarily inaccessible (e.g., temporary flooding) or data collection was missed because of a documented operational issue (e.g., unforeseen capacity issues that did not allow sampling in the temporal window), the sampling location will be considered missing for the given Reporting Period. These locations may be excluded from the stratum-level paired-difference calculation for that Reporting Period, and the statistical power of the analyses must be accordingly updated to reflect the altered sample size. These missing sampling points must be clearly documented at the relevant Verification. If data are missing for a sampling location during a second consecutive Reporting Period, the sampling point must be substituted.

Strata must be defined to support direct comparison of SOC stocks across Reporting Periods. Project Proponents may re-aggregate, split, or otherwise revise strata at a Reporting Period where this improves project-level precision, where field divergence under management has rendered the original stratification non-homogeneous, or where additional data has become available. Any change must:

• Preserve a documented mapping between the previous and revised stratification, so that prior measurements can be reconciled to the new design;
• Demonstrate that the revised design does not selectively exclude areas of expected SOC loss; and
• Be reported in the Monitoring Report and approved at verification.

Project-specific strata, their areas, the sampling locations within each, and any revisions across Reporting Periods must be reported as an annex to project documentation at every verification.

The requirements of this Section govern the design of the sampling campaign, specifically, whether the campaign is statistically capable of distinguishing a real change in SOC stocks from sampling noise. They do not determine the size of the uncertainty discount applied to issued credits, which is set separately under Section 9.1.1. The sampling design must be capable of detecting a project-level minimum detectable difference (MDD) in SOC stocks at the 90% confidence level with 90% power over the first Reporting Period. The MDD must be set prior to the first sampling campaign, must not exceed the expected project-level SOC stock change over that period, and must be approved at validation.

The power requirement applies at the primary quantification unit level, not stratum-by-stratum. Where a stratified design is used, the total number of samples must be allocated across strata to meet the project-level MDD using optimal (Neyman) allocation based on stratum area and within-stratum variance, or an equivalent allocation rule documented in the Project Monitoring Plan.

Where ancillary variables (e.g., remote-sensed indices, terrain attributes, digital soil maps) are used either to inform the design or to support a regression or model-assisted estimator at the project level, the sample size calculation may incorporate the variance reduction expected from those covariates, provided the correlation has been quantified using project-area or comparable regional data and is reported transparently.

A minimum of three composite samples per stratum is required to support variance estimation. Strata containing fewer than three samples must be pooled with the most similar neighboring stratum for estimation, following a documented and reproducible pooling rule.

The MDD-based power analysis below uses the SD of location-level paired SOC stock changes (the differences between t0t_0 t0​ and the most recent sampling event at each location), because this directly reflects the noise in what the project is trying to detect. Cross-sectional SD in absolute SOC stocks is not used and is generally a poor predictor of paired-difference variance.

The minimum number of samples needed to detect a given MDD at the project level is calculated as:

$$MDD = \frac{S}{\sqrt{n}} \times  

    (t_{\alpha,\upsilon} + t_{\beta,\upsilon})$$

(Equation 13)

$$n \geq \left(\frac{S \times (t_{\alpha} + t_{\beta})}{MDD}\right)^2$$

(Equation 14)

Where:

$MDD$ is the minimum detectable difference in SOC stocks (t C ha$^{-1}$);

$S$ is the standard deviation of the SOC stock change at fixed sampling locations — i.e., the location-level differences $\Delta SOC_{loc} = SOC_{loc,t_n} - SOC_{loc,t_0}$ computed at each composite sampling location — pooled across strata weighted by stratum area (t C ha$^{-1}$). "Pooled" here refers to the statistical aggregation of within-stratum SDs into a project-level SD for the purposes of this power analysis, and does not imply any physical pooling of samples.

$n$ is the minimum number of samples required;

$\upsilon = n - 1$ is the degrees of freedom;

$t_{\alpha}$ is the two-sided critical value of the t-distribution at significance level $\alpha$. $\alpha$ must not exceed 0.05, meaning the sampling design must control the probability of falsely concluding a SOC change has occurred when none has, to no more than 5%.

$t_{\beta}$ is the one-sided quantile of the t-distribution corresponding to the probability of a Type II error $\beta$. $\beta$ must not exceed 0.10, meaning the sampling design must achieve at least 90% statistical power to detect a true SOC change of magnitude $MDD$ when one is present.

The within-stratum standard deviation $S$ used in the design-stage power analysis must be estimated using the most direct evidence reasonably available for the project area, in the following order of preference:

1. Project-area pre-sampling. A pre-sampling campaign within the project area, designed to characterize within-stratum SOC variability at the spatial scale at which the full sampling campaign will be conducted. Pre-sampling is the preferred source under all conditions and is the default where database- or literature-derived variance is otherwise the only available source (see paragraph below).
2. Empirical variance estimates from peer-reviewed literature. Within-stratum or within-field SOC variance values reported in peer-reviewed studies conducted in the project's ecoregion, on comparable soil types, and at comparable spatial scales. Where a range of values is reported across studies, the upper end of the range applicable to the project area must be used.
3. Gridded predictive products. Variance estimates derived from gridded SOC prediction products (e.g., SoilGrids, SSURGO, or comparable national or regional digital soil maps) may be used only subject to the following conditions:

• The Project Proponent must explicitly document that the variance reported by the product reflects between-pixel prediction uncertainty conditional on the product's covariate structure, and does not, on its own, capture within-pixel local SOC heterogeneity at the spatial scale at which field samples will be drawn.
• The variance estimate used in the power analysis must be inflated to account for the within-pixel heterogeneity component. The inflation factor must be documented and justified using either (i) regional empirical studies of within-field or within-pixel SOC variability for comparable soil and management contexts, or (ii) a default inflation of the standard deviation by a factor of $\sqrt{2}$ (i.e., a doubling of the variance) where (i) is unavailable. Any inflation factor below $\sqrt{2}$ on the standard deviation must be justified at validation against published evidence specific to the project's ecoregion.
• Where the gridded product reports prediction uncertainty at a coarser spatial resolution than the project's quantification units, the variance estimate must be additionally inflated, or rejected as a source, to reflect the resolution mismatch. Where the product does not report any estimate of prediction uncertainty, it may not be used as a variance source.

Where database- or literature-derived variance is the only design-stage source available, the Project Proponent must conduct a pilot pre-sampling round of at least 10 sampling locations per stratum, drawn under the same probabilistic design as will be used for the full sampling campaign, prior to finalizing the sample size and committing to the first sampling campaign. The variance observed in the pilot replaces the database- or literature-derived estimate in the power analysis. The pilot samples are not credit-relevant and may be drawn at reduced analytical cost (e.g., proximal sensing where validated under Section 9.1.2.9, or single-increment composited samples).

The initial power analysis remains provisional and must be updated at each subsequent re-sampling event using observed within-stratum variance from the preceding Reporting Period. Where the observed variance is materially higher than was assumed at the design stage, the consequences for crediting precision are absorbed through the uncertainty discount at credit issuance under Section 9.1.1; where the observed variance is materially higher than the design-stage power analysis can support at the proponent's chosen MDD, the sample size must be increased for the next Reporting Period.

Rarefaction analysis at each re-sampling event. Beginning at the first re-sampling event ($t_1$), Project Proponents must conduct a rarefaction analysis to demonstrate empirically that each stratum's sample count supports stable estimates of the mean SOC stock change. The procedure is:

1. For each stratum, compute the bootstrap mean of the location-level paired SOC stock difference $\Delta SOC_{loc} = SOC_{loc,t_n} - SOC_{loc,t_0}$ using progressively larger random subsets of the available sampling locations, starting from a minimum subset of 5 locations and increasing in increments of 1 (or 5, where the stratum contains > 50 locations) up to the full sample set.
2. Repeat each subsample size at least 200 times to generate a distribution of bootstrap means at each subsample size.
3. Identify the smallest sample size $n^*_j$ at which the width of the 95% confidence interval of the bootstrap mean changes by less than 5% of the overall observed range of $\Delta SOC_{loc}$ in the stratum, for all subsequent increments. The stratum's design sample size $n_j$ must equal or exceed $n^*_j$.

Where any stratum fails the rarefaction criterion (no $n^*_j \leq n_j$ exists), location resampling under §9.1.1 Step 1 is not permitted for that stratum at that event, and the parametric fallback under §9.1.1 applies. The Project Proponent must increase $n_j$ for the next Reporting Period to meet the rarefaction criterion.

SOC stocks and stock changes must be reported to a minimum depth of 30 cm across all sampling locations, or to bedrock or hardpan where soils are shallower than 30 cm. In addition, a deep-core subsample comprising at least 10% of the sampling locations within each stratum must be extended to a minimum depth of 100 cm, or to bedrock or hardpan where soils are shallower than 100 cm.

Deep-core locations must be selected using the same probabilistic design used for the broader sampling campaign in that stratum (Section 9.1.2.2), must be representative of the soil and management conditions of the stratum as a whole, and must be georeferenced for consistent re-sampling across Reporting Periods. A minimum of three deep-core locations per stratum is required; where 10% of stratum locations would yield fewer than three, three deep cores must be sampled.

Deep cores must be collected at the baseline campaign at $t_0$​ and at every subsequent true-up campaign at which field measurements of SOC and bulk density are taken. Under Approach 1, this means every resampling event. Under Approach 2, this means every remeasurement (true-up) campaign as defined in Section 9.1.3.2.4; deep cores are not required at interim modeled crediting events, where no field measurements are taken.

Soils must be sampled using two depth increments (0–15 cm and 15–30 cm) at all sampling locations. At deep-core locations, the same 0–15 cm and 15–30 cm increments must be sampled, plus two additional increments of 30–60 cm and 60–100 cm. Where soils at a deep-core location are shallower than 100 cm, the deepest increment must be reported to the sampled depth and documented. SOC content analysis may be performed on a single composited sample per increment provided soil mass data are recorded separately for each increment.

Projects must apply an ESM correction. ESM correction must be applied across the 0–15 cm and 15–30 cm increments using soil mass data from all sampling locations, and across the 30–60 cm and 60–100 cm increments using soil mass data from the deep-core subsample.

For sampling events post-project initiation, this may require extension (up to 10cm) to a further depth in order to capture the mass established at project initiation. The actual depth of collection must be recorded and reported, and the additional mineral mass below 100 cm must be included in the ESM correction calculation for the 60-100 cm increment. Where a 10 cm additional buffer does not capture an equivalent mass of soil compared to project initiation, the deepest increment must be reported on a fixed-depth basis and documented as such. Note that baseline measurements at project initiation do not require multiple depth increments for ESM purposes within a given depth band.

Stratum-level SOC stocks reported under Equation 16 must reflect the full 0–100 cm sampled profile (or to bedrock/hardpan where shallower), with the 0–30 cm component estimated from all sampling locations and the 30–100 cm component estimated from the deep-core subsample. The credited cumulative SOC stock change must be calculated on this basis.

Soil sampling must follow established best practices for field collection and laboratory processing. The following requirements apply:

• Both the intended and actual sampling point locations must be recorded and georeferenced.

• SOC content, oven-dry fine soil mass, and sample volume must be obtained from the same sample or from adjacent samples taken during the same sampling event, if the sample size is insufficient. Where multiple cores are combined into a single sample, all cores must be taken from the same depth and fully homogenised prior to subsampling for the different measurements and this must be disclosed in the documentation of the sampling procedures.

• All organic material (e.g., living plants, crop residue) must be cleared from the soil surface prior to sampling. This must be documented via a photograph taken after sample removal.

• The fine soil mass per unit area for each depth increment must be derived from the oven-dry mass of soil passing a 2 mm sieve (excluding gravel, stones, and plant material) and the sampled volume corresponding to that increment. Coarse material must be prevented from passing through a 2 mm sieve.

• For the purposes of Equation 16, the soil mineral mass per unit area for each depth increment must be derived from the fine soil mass per unit area, adjusted to exclude the mass of soil organic matter:

  $$M_{min,j,d,t_n} = M_{fine,j,d,t_n} \times \left(1 - 1.724 \times C_{j,d,t_n} \times 10^{-3}\right)$$
  

  (Equation 15)

  • Where:
    • $M_{fine,j,d,t_n}$ is the oven-dry fine soil mass per unit area in depth increment $d$ of stratum $j$ at sampling event $t_n$ (kg m$^{-2}$),
    • $C_{j,d,t_n}$ is the measured SOC concentration in that increment (g C kg$^{-1}$ fine soil),
    • and 1.724 is the van Bemmelen conversion factor from organic carbon mass to organic matter mass. Where Project Proponents have project-area-specific data on the carbon fraction of soil organic matter, that value may be substituted for 1.724 with justification at validation.

• Bulk density, where reported, is a derived quantity calculated as the oven-dry fine soil mass divided by the sampled volume. It is not used as an independent input to Equation 16.

• Drying and sieving procedures must follow laboratory-specific standard operating procedures (SOPs) and must be applied consistently across all samples throughout the project lifetime, including where there is a change in analytical laboratory. Sample processing procedures must be reported in detail, explicitly describing sieving and grinding procedures.

"Composite" in this Section refers to the physical homogenization of multiple cores collected at a single sampling location, prior to laboratory analysis. The composite is the analytical sample; the sampling location is the statistical unit. Pairwise SOC stock differences computed at the sampling-location level (the difference between the composite measurement at tnt_n tn​ and at t0t_0 t0​) are the basis for stratum-level variance estimation in §9.1.2.3. Project Proponents are not required to physically pool samples across sampling locations or across strata at any point.

Following sampling, samples may be temporarily stored on-site in a location protected from sunlight, humidity, and precipitation, with different soil materials kept separate. Soil samples must be shipped within five days, or stably stored (e.g., dried or refrigerated, but not frozen) in a way that will be maintained until analysis. Once shipped, samples must be stored under environmentally controlled conditions that minimize biological activity (e.g., dried or refrigerated, but not frozen) until analysis. The duration of refrigerated storage prior to analysis must not exceed three months.

The selected analytical laboratory must be listed as an approved analytical service provider for SOC measurements in accordance with national or international accreditation standards. The laboratory must hold ISO/IEC 17025 accreditation or operate under a documented equivalent quality assurance framework. Where an equivalent framework is relied upon, the Project Proponent must demonstrate equivalence to ISO/IEC 17025, addressing, at a minimum, method validation, measurement traceability, internal quality control, personnel competence, and proficiency testing, and this demonstration shall be reviewed and accepted at project validation.

All samples collected throughout the project lifetime should be analyzed by the same laboratory wherever practicable. A transition to a different laboratory is permitted, including in cases where the incumbent laboratory is no longer able or willing to process project samples, provided that:

• The change is justified in writing and documented for review at the subsequent verification;
• The replacement laboratory satisfies the accreditation requirements set out above; and
• A documented cross-calibration procedure is completed prior to the new laboratory's results being used for quantification. The cross-calibration must involve parallel analysis of a representative set of samples (including, where available, archived samples and certified reference materials) by both laboratories, with results compared against pre-defined acceptance criteria for bias and precision. Any systematic offset identified must be addressed through method alignment or a documented correction, and the cross-calibration report must be submitted as part of project documentation.

The selected laboratory must quantify and report analytical error statistics to the Project Proponent on a regular basis, derived from repeated analyses of the same sample and from analyses of certified reference materials. The laboratory must provide documentation of its internal quality control program, including:

• Use of certified soil reference materials with known SOC content (e.g., NIST SRM 2710, BCR-129, or an equivalent traceable certified reference material appropriate for the SOC content range of the project samples);
• Monitoring of variation in analysis against defined error thresholds; and
• Participation in external proficiency testing schemes (e.g., round-robin testing) or registration as a member of the Global Soil Laboratory Network (GLOSOLAN), or an equivalent nationally recognized program.

Quantification of analytical uncertainty for Monte Carlo propagation. For each batch of project samples, the analytical uncertainty distribution used to perturb SOC concentration measurements in the Monte Carlo simulation (§9.1.1) must be characterized as follows:

1. Systematic bias correction. Identify any systematic bias in the analytical run against the certified reference material(s), expressed as the mean signed difference between measured and certified SOC values. Where systematic bias exceeds the CRM's certification uncertainty, project sample measurements must be corrected for the bias and the residual uncertainty (after correction) documented in the Monitoring Report.
2. Random error characterization. The random analytical error must be characterized as the greater of:

• (i) the relative standard deviation (RSD) of laboratory duplicate analyses on project samples within the same analytical batch, expressed as a fraction of the measured value; or
  • (ii) the relative error derived from CRM recovery — i.e., the standard deviation of CRM measurements within the batch divided by the certified value.

3. Monte Carlo input distribution. The random error from Step 2 must be used to parameterize a mean-zero perturbation around each measured SOC concentration value within the Monte Carlo simulation under §9.1.1. The distribution may be assumed Gaussian provided the laboratory's quality-control data support this assumption; where the duplicate/CRM distribution is materially non-Gaussian, the empirical distribution must be used directly via inverse-transform sampling or a documented alternative.

A minimum of 5% of project samples within each analytical batch must be analyzed as laboratory duplicates. Where the batch size is small (< 20 samples), a minimum of three duplicates must be analyzed regardless of percentage. Where bulk density is measured separately from SOC concentration, analytical uncertainty must be characterized separately for each parameter and propagated jointly in the Monte Carlo simulation.

The Project Proponent must retain run-level analytical data for all project samples, including individual measurement results, replicate analyses, calibration records, reference material results, and associated QC flags, and ensure data is available for audit by Isometric or a VVB. Quality control documentation, together with any cross-calibration reports, must be submitted as an annex to project documentation at each verification.

SOC content must be measured using dry combustion (Dumas method) with known and reported measurement uncertainty from the specific analytical run (i.e., default values not accepted).

Walkley-Black (wet) oxidation and loss on ignition (LOI) are not permitted except where no other analytical method is available, in which case their use must be justified in the Project Monitoring Plan and approved at validation. Approval may require a subset of samples to be sent for analysis using dry combustion for calibration and validation of these methods. Project Proponents must document the known limitations of these methods and apply appropriate uncertainty adjustments to the resulting SOC stock estimates.

In-situ proximal sensing techniques may be used as alternatives or complements to laboratory dry combustion analysis for the quantification of SOC. The following techniques are permitted under this Protocol, subject to the requirements below:

• Infrared spectroscopy, including near infrared (NIR), visible near infrared (Vis-NIR), and mid-infrared spectroscopy (MIR);

• Laser-induced breakdown spectroscopy (LIBS); and

• Inelastic neutron scattering (INS, also known as neutron-stimulated gamma ray analysis or spectroscopy).

Prior to use, Project Proponents must demonstrate, in agreement with Isometric, that the selected technique is equivalent in accuracy and reliability to dry combustion analysis for the soil types, moisture conditions, and SOC content ranges present in the project area.

This demonstration must be grounded in published peer-reviewed scientific literature and must be submitted for review and approval by Isometric prior to project validation. Techniques not yet supported by sufficient published evidence of equivalence to approved measurement methods will not be permitted.

All proximal sensing instruments must be calibrated against reference samples from the project area with SOC content determined by dry combustion prior to deployment and at regular intervals throughout the project lifetime. Calibration procedures must follow methods described in published peer-reviewed literature and must be conducted in consultation with Isometric. Prior to use for credit-relevant measurements, instruments and calibration models must be validated against independent reference samples not used in model development, with validation procedures and acceptance thresholds consistent with those reported in peer-reviewed literature and approved by Isometric.

Uncertainty associated with proximal sensing measurements must be quantified and propagated through the entire removal calculation in accordance with Section 7.6 of the Improved Soil Management Protocol. Where proximal sensing uncertainty exceeds that achievable through laboratory dry combustion, the more conservative estimate must be used for credit issuance. A subset of samples measured by proximal sensing must be independently verified by laboratory dry combustion at each Reporting Period.

SOC stocks within each stratum must be calculated on an equivalent soil mineral mass (ESM) basis. Use of soil mineral mass, rather than fine soil mass, ensures that the reference quantity against which SOC stocks are normalized is itself insensitive to changes in soil organic matter content driven by project activities. Sampling events are denoted $t_n$, where $t_0$ is the baseline measurement at project initiation and $t_1$, $t_2$, ... are subsequent resampling events.

The ESM-corrected SOC stock density for stratum $j$ at sampling event $t_n$ is calculated using a fixed-increment formulation across the depth increments required under Section 9.1.2.4:

$$SOC_{j,t_n} = \sum_{d=1}^{D} \left( C_{j,d,t_n} \times M_{min,ref,j,d} \right) \times \frac{1}{100}$$

(Equation 16)

Where:

• $SOC_{j,t_n}$ is the ESM-corrected SOC stock density in stratum $j$ at sampling event $t_n$ (t C ha$^{-1}$);
• $C_{j,d,t_n}$ is the measured SOC content in depth increment $d$ of stratum $j$ at sampling event $t_n$ (g C kg$^{-1}$ fine soil);
• $M_{min,ref,j,d}$ is the reference soil mineral mass for depth increment $d$ in stratum $j$, defined as the minimum soil mineral mass per unit area observed across all sampling events for that increment (kg m$^{-2}$), calculated in accordance with Section 9.1.2.5;
• $D$ is the total number of depth increments required under Section 9.1.2.4 (typically four: 0–15 cm, 15–30 cm, 30–60 cm, 60–100 cm). Where soils within stratum $j$ are shallower than 100 cm, $D$ is reduced accordingly and the deepest increment is reported to the sampled depth;
• $\frac{1}{100}$ converts units from g C m⁻² to t C ha⁻¹.

Differential sampling intensity by depth. Consistent with Section 9.1.2.4, the parameters in Equation 16 are estimated from the sampling locations at which the relevant depth increment was sampled:

• For the 0–15 cm and 15–30 cm depth increments, the parameters $C_{j,d,t_n}$ and $M_{min,ref,j,d}$ are estimated using the full set of sampling locations within stratum $j$.
• For the 30–60 cm and 60–100 cm depth increments, the same parameters are estimated using the deep-core subsample only within stratum $j$ (i.e., the ≥ 10% of sampling locations sampled to 100 cm or to bedrock/hardpan in accordance with Section 9.1.2.4).

Cross-event minimization under the ESM correction is applied within each population (i.e., separately for the 0–15 cm / 15–30 cm increments measured across all locations, and for the 30–60 cm / 60–100 cm increments measured across deep-core locations), not pooled across populations.

Where only a single depth increment has been sampled at $t_0$ (e.g., where the deep-core subsample minimum has not yet been satisfied), $M_{min,ref,j,d}$ for that increment must be set equal to $M_{min,j,d,t_0}$ such that the ESM correction ratio equals unity at baseline. The ESM correction takes full effect at subsequent sampling events $t_1, t_2, ...$, once the cross-event minimum can be calculated against more than one observation.

The variance terms used in Section 9.1.1 (Uncertainty Propagation) must reflect the differential sample size by depth increment and must be propagated separately for the 0–30 cm and 30–100 cm components of the SOC stock estimate before being combined at the stratum level.

Under this approach, an approved biogeochemical model (see Section 9.1.3.2.1) is used to estimate SOC stock changes between resampling campaigns based on measured initial SOC stocks, implemented land management practice changes, soil characteriztics, and climatic conditions within each quantification unit.

Direct measurement of SOC stocks is required at a minimum of every five years. Remeasurement data must be used to re-estimate model prediction error and recalibrate the model against observed conditions at each Reporting Period (true-up procedure, see Section 9.1.3.2.4).

Model validation must be conducted in accordance with the requirements set out in Section 9.1.3.2.3. The IME is responsible for approving all datasets used for calibration and validation purposes.

Modeled carbon removal estimates carry inherently greater uncertainty than directly measured values, arising from model error, parameter uncertainty, temporal extrapolation, and the inability to directly observe the baseline scenario. A tiered discount framework therefore applies:

• Interim crediting events (between resampling campaigns) --- an uncertainty discount of 1.2 SD of the estimated project effect must be applied to modeled carbon removal estimates. This reflects the greater uncertainty inherent in modeled interim estimates and provides appropriate conservatism against systematic overcrediting where model uncertainty may be underestimated.
• True-up events (resampling campaigns) --- the uncertainty discount may be reduced to the standard 1.0 SD of the estimated cumulative project effect, where all of the following conditions are met: cumulative SOC stock changes have been directly measured through a resampling campaign in accordance with Section 9.1.2; the cumulative accounting position has been formally reconciled against all credits previously issued; and the model's cumulative performance has been assessed through the true-up procedure in Section 9.1.3.2.4 with no systematic overprediction detected.

Projects that defer or delay resampling campaigns beyond the scheduled interval may not apply the 1.0 SD discount until a compliant true-up has been completed.

Any model used to contribute to the quantification of net CO₂e removal under this Protocol must be demonstrated to be well-validated and skillful for the purpose for which it is used, including the relevant soil types, land management practices, climatic conditions, and geographic context of the project area. Recommended biogeochemical models include established process-based models such as DayCent, RothC, and CENTURY, as well as other models that meet the following eligibility criteria. However, even recommended models must be demonstrated to be fit for purpose in the context of the Project according to the below criteria.

Model eligibility must be demonstrated through one of the following two pathways, submitted in the PDD and approved by Isometric prior to Validation:

• Established models

  • The model has a track record of use in science, industry, or government applications, demonstrated through multiple peer-reviewed publications reporting its application to SOC stock change estimation under land management conditions comparable to those of the project area.

  • The model must be demonstrably relevant to the ecoregion, soil types, and land management practices present in the project.

• Newly developed models

  • Where a model does not yet have an established peer-reviewed track record, it must be validated against reputable independent data sources prior to use.

  • Validation data must comprise quality-controlled in-situ SOC measurements and publicly available datasets adhering to FAIR (Findable, Accessible, Interoperable, and Reusable) principles.

  • Sufficient validation data and results must be submitted with the PDD and approved by Isometric prior to validation.

Model eligibility is context-specific. A model that is eligible for one project area is not automatically eligible for a project in a different ecoregion, soil type, or management context. Where the same model is to be used both to predict the project scenario and to estimate the counterfactual under Section 9.2.2, eligibility must be demonstrated separately for the project management practices and the baseline management practices, in line with Section 9.1.3.2.3.

The selected model must be capable of simulating SOC dynamics across the depth profile required under Section 9.1.2.4, including the 30–100 cm profile. Models that simulate only topsoil dynamics (commonly 0–30 cm or shallower) should be coupled with a complementary mechanism for simulating subsoil SOC dynamics that is itself separately validated under Section 9.1.3.2.3.

Model calibration and parameterization must be fully documented and reproducible. Project Proponents must provide sufficient information in the PDD for an independent third party to replicate the model setup and obtain equivalent outputs from the same inputs. At a minimum, the following must be documented:

• All parameter values used, including their source (e.g., published literature, national databases, site-specific measurements, or model defaults);
  • Project Proponents must clearly specify which model parameters were subject to calibration
• The version of the model used, including any modifications made to the base model code or configuration;
• The procedure for model calibration reported in enough detail to enable reproducibility;
• All input data used to initialize and run the model, including measured initial SOC stocks, soil physical and chemical properties, climate data, and land management practice records; and
• Any assumptions made where site-specific data were unavailable, including justification for the values adopted.

All data sources used in parameterization must be available to Isometric and the VVB. Where proprietary data sources are used, Project Proponents must demonstrate that equivalent publicly available data were not available and must provide sufficient metadata to allow independent assessment of data quality.

Digital soil maps may be used for model initialization with approval from Isometric.

Prior to use for credit-relevant quantification, the selected biogeochemical model must be validated for the specific conditions of the project area. Validation results and supporting data must be submitted in the PDD and are subject to review and approval by Isometric at project validation.

Data points used for model parameterization or calibration must not be used for validation. This prohibition applies to all data sources used in model development, including literature-derived values used at initial parameterization and any project-collected data subsequently used for recalibration under the true-up procedure. A minimum of 20% of all available representative data points must be withheld from parameterization and reserved exclusively for validation. Representative means that the withheld data points must span the full range of SOC stock values, soil types, climatic conditions, and land management practices present in the project area. A validation subset that satisfies the 20% threshold but clusters at one end of the observed distribution does not meet this requirement.

The validation dataset must:

• Include representative coverage of SOC stock values across the full 0–100 cm depth profile required under Section 9.1.2.4, including the 30–100 cm subsoil component;
• Be drawn from the same ecoregion and encompass soil types, climatic conditions, and land management practices comparable to those of the project area; and
• Include the range of management practice changes implemented under the project, so that model performance under project-relevant conditions can be directly assessed.
• Where the model is to be used to estimate the counterfactual under Section 9.2.2, additionally include observations representative of the baseline management practices that will be input to the counterfactual simulation, so that model performance under baseline-relevant conditions can be directly assessed.

Where the validation dataset is used to support a modeled counterfactual under Section 9.2.2, model performance must be reported separately for project-like and baseline-like management contexts and for project-area-relevant and project-area-non-relevant climatic contexts. The $R^2$ performance threshold and bias requirements set out in this section apply independently to each project-like × climatic-coverage subset and to each baseline-like × climatic-coverage subset; where they are not met for the baseline-like × project-area-relevant climatic-coverage subset specifically, the model is not eligible to be used for counterfactual estimation under Section 9.2.2, regardless of its performance under other subsets.

Climatic coverage of the validation dataset must be reassessed at each Reporting Period against the climatic conditions actually experienced by the project area in that period. Where the validation dataset does not include observations spanning the climatic conditions actually experienced — defined as fewer than three validation points within the climatic envelope of the Reporting Period under each of project-like and baseline-like management — the climatic-coverage gap must be documented and triggers the additional consequences set out in Section 9.1.3.2.4.

Based on assessment against the validation dataset, the model must demonstrate the following:

• $R^2$, calculated as the coefficient of determination $(1 - \frac{RSS}{TSS})$, must be greater than 0. A 90% confidence interval for $R^2$ must be calculated and must exclude 0, demonstrating statistically significant predictive skill. $R^2$ must not be calculated as the coefficient of determination of a linear regression of actual against predicted values, as this can produce misleadingly high values for biased models.
• Where model-based error propagation is used, the coverage of model predictions must be evaluated — the proportion of observed values falling within the model's predicted uncertainty intervals must be assessed and reported. This requirement does not apply where model-assisted or analytic error propagation is used.
• No systematic bias in predictions across the range of SOC stock values observed in the project area. Where bias is present, it must be characterized, documented, and corrected prior to use.

The following statistics must be calculated and reported as documentation requirements at validation, but do not constitute enforceable performance thresholds: root mean square error (RMSE), reported in t C ha$^{-1}$; and mean bias error (MBE), reported in t C ha$^{-1}$, as the primary diagnostic for systematic overprediction or underprediction.

Residual prediction errors must be randomly distributed with respect to soil type, management practice, and time. Structured residuals indicate that model errors are driven by factors not captured in the model and must be investigated and resolved before the model is approved for use.

Where a model does not meet the $R^2$ performance threshold or exhibits systematic bias that cannot be corrected, it must be reparameterized or recalibrated before resubmission for Isometric approval. Data points previously used for validation may not be reused for parameterization or recalibration.

Where the Project Proponent can demonstrate that representative in-situ data are not reasonably available for the specific combination of soil, climatic, and management conditions of the project area, model performance over a generalized parameter space encapsulating the project conditions may be considered for validation, subject to consultation with and approval by Isometric. The Project Proponent must document the data scarcity, characterize the parameter space against which generalized validation is conducted, and demonstrate that this space is sufficiently broad and representative to provide reasonable confidence in model performance under project conditions. The performance thresholds set out in this Section apply unchanged to the generalized validation.

The validation dataset is constructed by withholding a minimum of 20% of representative data points from a broader corpus of available evidence. The integrity of this withholding process depends on the corpus itself being complete, since selective construction of the corpus before the 20% withholding is applied is functionally equivalent to selective construction of the validation set. Project Proponents must therefore construct the validation data corpus to the following completeness standard.

The corpus must include all peer-reviewed studies that satisfy each of the following criteria:

1. Published in the peer-reviewed literature within the 25 years preceding the validation date;
2. Drawn from the project's ecoregion, defined consistently with the management-coverage and climate-coverage requirements set out elsewhere in this Section;
3. Reporting SOC stocks (not solely SOC concentration) at depth profiles compatible with the project's reported sampling depths under Section 9.1.2.4, including, where applicable, the deep-core profile to 100 cm; and
4. Reporting SOC stocks under management practices identifiable as project-like or baseline-like under the partition defined elsewhere in this Section.

The Project Proponent must document the literature search procedure used to identify candidate studies, including the databases queried (at minimum: Web of Science, Scopus, AGRIS, and Google Scholar), the search terms used, the date of the search, and the total number of studies identified before any exclusion is applied.

Studies satisfying the mandatory inclusion criteria may be excluded from the corpus only on the basis of one or more of the following grounds, with each exclusion documented individually at the study level and subject to scrutiny at validation:

1. Methodological incompatibility — the study uses an SOC measurement method (e.g., wet oxidation with no dry-combustion comparator, or a non-comparable depth-aggregation procedure) that cannot be reconciled with the project's measurement framework;
2. Documented data quality issues — the study has been subject to a published correction, retraction, or methodological criticism that materially undermines the reliability of its SOC stock measurements;
3. Soil-type incompatibility — the study covers a soil type that is not present in the project area, where soil type is the primary control on the model's predictions;
4. Replacement by superseding study — the study has been superseded by a later peer-reviewed remeasurement of the same locations, in which case the later study replaces (rather than is added to) the earlier one in the corpus; or
5. Inaccessibility — the underlying data are not available to the Project Proponent through reasonable efforts, including direct request to the corresponding author. Where this exclusion is invoked, the Project Proponent must document the specific data-request actions taken and the response (or non-response) received.

Exclusion on grounds outside this list is not permitted. In particular, a study may not be excluded on the basis that its inclusion would widen the Monte Carlo input distributions or increase the uncertainty discount.

Changes to the corpus or to the validation subset after pre-registration require explicit Isometric approval and constitute a fresh validation event for which model performance metrics calculated prior to the change may not be relied upon.

Calibration and validation data points must be spatially independent. Spatial mixing of calibration and validation locations within the same continuous soil or bioclimatic neighborhood produces residual correlation between the two sets that systematically inflates reported model performance statistics relative to the model's true generalization skill. The data-separation rule set out above (no point used for both parameterization and validation) is necessary but not sufficient for this purpose; spatial separation is additionally required and must be implemented through one of the following two designs:

1. Distance-based separation. Calibration and validation points are drawn from the same dataset, but each validation point must lie beyond the empirical spatial autocorrelation range of the SOC residual variogram from the nearest calibration point. The autocorrelation range must be estimated from the project's available data (or, where insufficient project-area data are available, from regional comparable data) using a documented variogram fitting procedure, and is defined as the lag distance at which the empirical semivariogram reaches 95% of its sill value. Where the variogram is anisotropic, the longer of the two principal-axis ranges must be used.
2. Block-level hold-out. Entire spatial blocks — defined as project strata, quantification units, sub-regions, or other contiguous spatial units of meaningful agronomic or pedogenic homogeneity — are reserved for validation rather than individual points within shared blocks. Where this design is used, the minimum number of held-out blocks must be sufficient to support the validation framework's statistical-power requirements (which under the prior proposed redraft are evaluated separately for each project-like × climate-coverage and baseline-like × climate-coverage subset), and the held-out blocks must collectively span the management-coverage, climate-coverage, and distributional-distance criteria set out elsewhere in this Section.

Project Proponents must report at validation, for each bias-test subset under Section 9.1.3.2.4 (project-like, baseline-like, and the climate-coverage portions of each):

• The empirical SOC residual variogram or equivalent autocorrelation diagnostic, including the fitted autocorrelation range and 90% confidence interval on the range estimate;
• The minimum, median, and 5th-percentile distances between validation locations and the nearest calibration location;
• Where the distance-based separation design is used, evidence that the 5th-percentile distance exceeds the autocorrelation range; and
• Where the block-level hold-out design is used, the spatial definition of each held-out block, the rationale for the choice of block size, and a demonstration that no calibration-point neighbourhood (defined by the autocorrelation range) crosses a block boundary into a held-out block.

Where the project area is too compact to support distance-based separation beyond the autocorrelation range, or where the available validation dataset is too small to populate block-level hold-outs at the required statistical-power thresholds, within-project spatial blocking is not feasible and the Project Proponent must instead use regional validation data drawn from beyond the project boundary, with the regional dataset itself satisfying the management-coverage, climate-coverage, distributional-distance, and spatial-blocking requirements of this Section. The regional dataset must be drawn from the same ecoregion, must satisfy the comparability criteria set out elsewhere in this Section, and must include sufficient spatial extent for distance-based separation to be operationalised within it. Use of the fallback must be documented at validation, including a quantitative demonstration that within-project blocking is infeasible and a justification of the regional dataset's representativeness of the project area.

Where post-recalibration validation is conducted under Section 9.1.3.2.4 following a triggered bias test, the spatial-blocking requirements of this sub-section apply equally to the post-recalibration validation. New calibration points used in recalibration must be spatially separated from validation points under the same design as was used at initial validation. Where new calibration data are spatially correlated with existing validation data such that this requirement cannot be satisfied without restructuring the validation set, the Project Proponent must construct a fresh validation set under the spatial-blocking requirements before the recalibrated model may be used for credit-relevant quantification.

At a maximum interval of 5 years, the biogeochemical model must be updated using new SOC stock measurements collected during the resampling campaign in accordance with Section 9.1.2. The true-up procedure serves three purposes: integrating new observational evidence into the model's cumulative data record; reassessing model error and uncertainty; and detecting and correcting systematic bias before it propagates into subsequent credit estimates.

New field measurement data must be added to the full cumulative dataset comprising all prior literature-derived values and project-collected data used in parameterization and validation. New data points must first be used for model error reassessment before any portion is considered for recalibration, the same data points may not be used simultaneously for both purposes within the same Reporting Period.

At each true-up event, the deep-core subsample collected under Section 9.1.2.4 must be used alongside the standard 0–30 cm sampling to (i) update model error statistics separately for the 0–30 cm and 30–100 cm components of the model prediction, and (ii) test for systematic bias in the modeled subsoil profile. The MBE-based bias test (10% of mean observed SOC stock change) applies independently to each profile component. Where systematic overprediction is detected in either component, the consequences set out in this section apply.

Updated error statistics must be calculated against the new observations and reported in the Monitoring Report at verification. At minimum the following must be reported, calculated separately for the project-like and baseline-like subsets of the cumulative validation dataset (as defined in Section 9.1.3.2.3) and for each subset's project-area-relevant climatic-coverage portion: $R^2$ calculated as $1 - \frac{RSS}{TSS}$, with a 90% confidence interval that must exclude 0; RMSE in t C ha$^{-1}$; and MBE in t C ha$^{-1}$, calculated as the mean of (predicted − observed) across all validation points in each subset, reported alongside its standard error and a one-sided 90% upper confidence bound on $|MBE|$. The pooled statistics for the cumulative validation dataset as a whole must additionally be reported. Updated error statistics must be used to revise the Monte Carlo simulation input probability distributions in accordance with Section 9.1.1. Where the true-up reveals greater model uncertainty than was previously characterized, the higher uncertainty estimate must be applied.

Three parallel bias tests apply at each true-up. Each test combines a statistical-significance criterion (to control false positives and false negatives at small or noisy validation sample sizes) with a residual minimum effect-size criterion (to prevent triggering on statistically detectable but practically immaterial bias). Any single test, if triggered, requires the mandatory consequences set out below.

Pooled bias test. The pooled-MBE test fires where both of the following are satisfied:

• The lower bound of a one-sided 90% confidence interval on MBE on the cumulative validation dataset as a whole exceeds zero — i.e., the test rejects the null hypothesis of no overprediction at the 10% one-sided significance level; and
• The point estimate of MBE on the cumulative validation dataset exceeds 5% of the mean observed SOC stock change in the project area across the cumulative dataset.

Where the test fires with a positive MBE point estimate (the model over-predicts), this constitutes systematic pooled overprediction. A negative MBE point estimate, or failure of either criterion, does not trigger the mandatory consequences but must be documented and investigated where statistically significant.

Differential bias test (management-conditional). The differential test fires where both of the following are satisfied:

• The lower bound of a one-sided 90% confidence interval on $(MBE_{proj} - MBE_{base})$ exceeds zero, where $MBE_{proj}$ and $MBE_{base}$ are the point estimates on the project-like and baseline-like subsets of the cumulative validation dataset respectively, and the confidence interval is calculated using a Welch-style standard-error formulation that accommodates unequal sample sizes and unequal within-subset variances between the two subsets; and
• The point estimate $(MBE_{proj} - MBE_{base})$ exceeds 5% of the mean observed SOC stock change in the project area across the cumulative dataset.

Where the test fires, the model is biased toward over-estimating the project effect (over-crediting). A negative point estimate, or failure of either criterion, does not trigger the mandatory consequences but must be documented and investigated where statistically significant.

Differential bias test (climate-conditional). The climate-conditional differential test fires where both of the following are satisfied:

• The lower bound of a one-sided 90% confidence interval on $(MBE_{proj,C} - MBE_{base,C})$ exceeds zero, where $MBE_{proj,C}$ and $MBE_{base,C}$ are the point estimates on the project-like and baseline-like subsets of the cumulative validation dataset restricted to the climatic-coverage portion overlapping with the climatic conditions actually experienced by the project area during the Reporting Period, calculated using the same Welch-style standard-error formulation as the management-conditional test; and
• The point estimate $(MBE_{proj,C} - MBE_{base,C})$ exceeds 5% of the mean observed SOC stock change in the project area across the cumulative dataset.

Where either subset under the climate-conditional test contains fewer than three validation points within the relevant climatic-coverage portion, the test is not statistically defined, and the climatic-coverage gap consequences set out below apply in lieu of the test result.

Mandatory consequences where any test fires. Where any of the three bias tests fires, the following mandatory steps apply in sequence:

1. Credits for the current Reporting Period must be calculated using observed SOC stock changes from field measurements rather than modeled estimates for the project-area component of the project effect. Where the management-conditional or climate-conditional differential test fires, the modeled counterfactual estimate for the current Reporting Period must additionally be adjusted by subtracting $MBE_{base}$ (or $MBE_{base,C}$ where the climate-conditional test fires) from the modeled counterfactual SOC stock change. The bias-correction adjustment must be capped at the lesser of (i) 25% of the unadjusted modeled counterfactual stock change for the period, or (ii) the absolute value of the bias point estimate from the relevant subset, with the cap and any application of it documented in the Monitoring Report. The combined project-effect estimate is then calculated as the difference between the field-measured project SOC stock change and the bias-corrected modeled counterfactual.
2. The model must be recalibrated and the results reported in accordance with Section 9.1.3.2.2 prior to the next Reporting Period. Recalibration must specifically target the source(s) of the bias identified by the test(s) that triggered. Recalibration must restore each fired test to non-firing status (both criteria failing on the post-recalibration validation dataset), evidenced through subset-by-subset performance reporting under Section 9.1.3.2.3 as updated by the recalibration.
3. The source of the bias must be investigated, documented, and reported to Isometric. Where the source cannot be identified or corrected, continued use of the model must be approved by Isometric before the next Reporting Period commences.

Climatic-coverage gap consequences. Where the climate-conditional differential test cannot be statistically defined for a Reporting Period because the validation dataset contains fewer than three validation points within one or both subsets restricted to the climatic conditions actually experienced, the consequences are as follows:

1. For the current Reporting Period, the modeled counterfactual estimate must be subjected to a default conservative adjustment: the modeled counterfactual SOC stock change must be increased by an amount equal to the larger of (i) the absolute value of $MBE_{base}$ on the broader baseline-like subset (i.e., the management-conditional, not climate-conditional, point estimate), or (ii) 5% of the mean observed SOC stock change in the project area across the cumulative dataset. The combined project-effect estimate is then calculated as the difference between the modeled (or measured, depending on Reporting Period type) project SOC stock change and the adjusted modeled counterfactual. The adjustment must be applied without the bias-correction cap that applies under the bias-test consequences above.
2. The Project Proponent must, prior to the next Reporting Period, expand the validation dataset to remedy the climatic-coverage gap. Where the climatic conditions experienced are unprecedented in the available validation evidence base for the project's ecoregion, this requirement may be satisfied by acceptance from Isometric of a documented gap, with the conservative adjustment under (1) continuing to apply at each subsequent Reporting Period until the gap is closed.

Project Proponents may elect to recalibrate model parameters at any true-up event using the updated cumulative dataset, subject to the following requirements:

• Recalibration must use the full cumulative dataset of literature-derived and project-collected data, excluding all data points reserved for validation;

• A minimum of 20% of all available representative data points in the updated cumulative dataset must be withheld from recalibration and used for post-recalibration validation, consistent with the data separation requirements of Section 9.1.3.2.3. A validation subset that satisfies the 20% threshold but clusters at one end of the observed distribution does not meet this requirement;

• Post-recalibration model performance must satisfy the $R^2$ performance threshold of Section 9.1.3.2.3, that is, $R^2$ greater than 0 with a 90% confidence interval excluding 0, and must demonstrate no systematic bias before the recalibrated model is used for credit-relevant quantification;

• All recalibration inputs, revised parameter values, updated validation results, evidence that the 20% representative validation split requirement has been satisfied, updated performance statistics, and revised Monte Carlo input probability distributions must be submitted to Isometric and approved prior to use of the recalibrated model in the next Reporting Period.

Data points previously used for validation may not be reused for parameterization or recalibration at any point in the project lifetime.

Remeasurement serves model calibration and true-up purposes rather than primary quantification. The following requirements apply to all resampling campaigns conducted under this approach:

• The QA/QC, georeferencing, seasonal consistency, and organic amendment timing requirements set out in Section 9.1.2.1 apply equally to remeasurement campaigns.

• The sampling design and stratification requirements of Section 9.1.2.2 apply equally to remeasurement.

• Stratification for remeasurement must be consistent with the stratification used for initial baseline measurement and for model parameterization, to ensure that remeasurement data are directly comparable to modeled outputs at the stratum level.

• The power analysis requirements of Section 9.1.2.3 apply to remeasurement campaigns. The $MDD$ must be set at a level sufficient to detect the model prediction error identified at initial validation, in addition to the expected project-induced SOC stock change.

• At true-up events, field measurements of bulk density must be collected during the resampling campaign and used to apply an ESM correction to the model output, consistent with the ESM approach selected under Section 9.1.2.4. The treatment of bulk density differs by event type:

  • Interim crediting events; the model output is used on a fixed-depth basis without an ESM correction, using the reference bulk density from baseline measurements ($BD_{j,ref}$). The absence of an ESM correction contributes to the application of the 1.2~SD uncertainty discount at interim events.

  • True-up events: measured bulk density data from the resampling campaign must be used to apply a full ESM correction to the model output, consistent with the approach described in Section 9.1.2.4.

The requirements of Section 9.1.2 and all subsections apply equally to remeasurement under Approach 2. In-situ proximal sensing (Section 9.1.2.9) may also be used for remeasurement, subject to the same requirements.

For projects where the model produces SOC predictions on a mineral-mass-equivalent basis (i.e., the ESM correction is part of the model formulation and prediction), the post hoc ESM correction in Equations 17 and 18 does not apply. In such cases, the model output can be used directly in Equation 12. Project Proponents electing this approach must i) clearly document this as being the case in the model's output within the model description in the PDD; ii) demonstrate the model's internal reference mass is consistent with, or more conservative than, the reference mineral mass defined in Section 9.1.2.10; and iii) ensure that model validation is conducted against observations on the same mass basis s the model output.

At interim crediting events, where no field measurements are available, the model output is used on a fixed-increment, mineral-mass basis without an ESM correction across events (the ESM correction is applied only at true-up events, where measurements are available to update the reference mineral mass). The reference soil mineral mass at interim events is taken from the baseline ($t_0$) measurements. The stratum-level modeled SOC stock density at modeled time step $t_n$ is:

$$SOC^{model}_{j,t_n} = \sum_{d=1}^{D} \left( \hat{C}_{j,d,t_n} \times M_{min,ref,j,d}^{(t_0)} \right) \times \frac{1}{100}$$

(Equation 17)

Where:

• $SOC^{model}_{j,t_n}$ is the modeled SOC stock density in stratum $j$ at modeled time step $t_n$, on a mineral-mass basis (t C ha$^{-1}$);
• $\hat{C}_{j,d,t_n}$ is the modeled SOC content in depth increment $d$ of stratum $j$ at time $t_n$ (g C kg$^{-1}$ fine soil);
• $M_{min,ref,j,d}^{(t_0)}$ is the baseline reference soil mineral mass for depth increment $d$ in stratum $j$ at $t_0$, calculated in accordance with Section 9.1.2.5 (kg m$^{-2}$);
• $D$ is the total number of depth increments required under Section 9.1.2.4 (typically four: 0–15 cm, 15–30 cm, 30–60 cm, 60–100 cm). Where soils within stratum $j$ are shallower than 100 cm, $D$ is reduced accordingly and the deepest increment is reported to the sampled depth;
• $\frac{1}{100}$ converts units from g C m$^{-2}$ to t C ha$^{-1}$.