Glossary

Key terms used in forest carbon credit assessment, counterfactual baseline analysis, geospatial foundation models, and carbon market integrity.

A5 Grid

A global equal-area grid system based on pentagonal cells. Unlike hexagonal grids (such as H3), which distort in area at different latitudes, pentagons maintain consistent area everywhere on earth. Pentagons also eliminate the continuous reprojection required by pixel-based raster workflows, where every analysis in a new location requires transforming data to a different coordinate system.

Additionality

The requirement that a carbon project demonstrates its climate benefits would not have occurred without the financial incentive from carbon credit sales. If a project protects a forest that was never under threat of deforestation, the climate benefit is not additional, and credits issued against it do not represent real emissions reductions.

AlphaEarth

A geospatial foundation model and embeddings dataset used by belian.earth in monitoring-period analysis of forest carbon project areas and reference areas.

ARR (Afforestation, Reforestation and Revegetation)

Carbon project activities that increase forest cover or biomass by establishing or restoring forest on land that is currently unforested or degraded. ARR projects generate removal credits, distinct from REDD+ activities which avoid emissions from existing forests. Verra's VM0047 is the leading ARR methodology in the voluntary carbon market.

Article 6 (Paris Agreement)

The article of the Paris Agreement that governs international cooperation on emissions reductions, including the transfer of mitigation outcomes between countries (ITMOs) and a centralised crediting mechanism overseen by the UNFCCC. Article 6 is the principal mechanism enabling sovereign carbon trading and links voluntary market activity to nationally determined contributions.

Article 6.2

The Paris Agreement mechanism that enables bilateral cooperation between countries on emissions reductions through Internationally Transferred Mitigation Outcomes (ITMOs). Article 6.2 allows countries to trade emissions reductions directly under bilateral arrangements that they govern themselves, subject to UN reporting requirements.

Article 6.4

The Paris Agreement mechanism establishing a UN-supervised crediting mechanism for international emissions reductions. Article 6.4 replaces the Kyoto Protocol's Clean Development Mechanism and is overseen by a UNFCCC Supervisory Body that approves methodologies and authorises credit issuance.

Avoided Deforestation

A category of carbon credit project that generates credits by preventing deforestation that would otherwise have occurred. The credibility of these credits depends entirely on the accuracy of the counterfactual baseline: the estimate of how much deforestation would have happened without the project.

Avoided Planned Deforestation

A category of REDD+ carbon project that prevents deforestation that was officially planned or licensed, for example forest clearance under a concession or government land conversion programme. Distinct from avoided unplanned deforestation.

Avoided Unplanned Deforestation

A category of REDD+ carbon project that prevents deforestation which would otherwise have occurred from unplanned drivers such as land conversion pressure, encroachment, or smallholder expansion. The most common form of REDD+ project, and the type for which counterfactual baseline science is most consequential.

Biomass Estimation

Biomass estimation involves quantifying the amount of living plant material, and therefore stored carbon, in a tree or forest. At the individual tree level, biomass is typically measured in kilograms of carbon. At the forest level, individual tree estimates are aggregated to give tonnes of carbon per hectare, often reported as tonnes of CO₂ equivalent per hectare. Methods range from field-based plot measurements to satellite-derived estimates. Approaches that combine remote sensing with local field plot data can produce locally calibrated models rather than relying on a single global biomass product.

Causal Inference

Statistical methods designed to estimate cause-and-effect relationships rather than mere correlations. In carbon markets, causal inference methods like synthetic control and pixel matching are used to construct credible counterfactual scenarios, answering the question of what would have happened without a carbon project.

CCI Biomass

The European Space Agency Climate Change Initiative biomass dataset, providing global maps of aboveground biomass derived from satellite remote sensing. Used in forest carbon assessment as one input among several for baseline biomass estimation.

CORSIA

The Carbon Offsetting and Reduction Scheme for International Aviation, the International Civil Aviation Organization's compliance market for the aviation sector. CORSIA requires airlines to offset growth in international flight emissions above a baseline, and the credits eligible under CORSIA must meet integrity criteria set by ICAO's Technical Advisory Body.

Counterfactual Baseline

An estimate of what would have happened to a forest area without a carbon project intervention. Unlike historical baselines that project past deforestation trends forward, counterfactual baselines compare project areas to similar areas that represent what would have happened in the absence of the conservation project. These comparison areas can be selected manually, by geographic proximity, or by using AI and causal inference methods to identify areas with genuinely similar characteristics. The resulting baseline provides a more credible estimate of the unprotected scenario than historical projection alone, because a causal inference approach will inherently capture unseen confounders not represented in the historic time series.

DMRV (Digital MRV)

Digital Monitoring, Reporting, and Verification. The use of remote sensing, AI, and automated systems to monitor carbon project performance, replacing or supplementing manual field-based verification. Enables more frequent, scalable, and transparent assessment of carbon projects.

Donor Pool

The set of candidate reference areas from which a synthetic control or matching method selects comparisons. Poor donor pool selection, such as including areas from different ecoregions or climatic zones, can undermine the entire counterfactual analysis.

Dynamic Baseline

Traditional baselines are fixed at the point of intervention start and not reevaluated for years thereafter. A dynamic baseline seeks to reevaluate on more regular intervals, using continuous earth observation data or detailed activity data from the ground rather than fixed assumptions. Dynamic baselines respond to real-world changes in deforestation pressure, land use policy, and economic conditions, making them more credible for continuous long-term carbon accounting.

Embedding

A vector representation of a place, produced by a foundation model, that encodes land surface characteristics in a form that can be compared mathematically. Two embeddings can be compared with a distance metric to assess how similar two locations are across the dimensions the model has learned to represent.

FARL (Forest Reference Activity Level)

The jurisdictional baseline used in REDD+ accounting, representing the reference rate of forest carbon emissions or removals for a country or sub-national region. FARLs are submitted to the UNFCCC and provide the level against which performance is measured under jurisdictional REDD+ frameworks.

Foundation Model

A large neural network, typically transformer-based, trained on broad data to learn general-purpose representations that can be reused across many downstream tasks. In earth observation, foundation models learn from billions of satellite observations to produce embeddings of the land surface that support biomass estimation, land-cover mapping, similarity matching, and counterfactual baseline construction.

Foundation Model Embeddings (Geospatial)

Numerical representations (vectors) of landscape characteristics extracted from large AI models pre-trained on satellite imagery. Each location on earth gets a vector that encodes information about its vegetation, terrain, land use, and seasonal patterns. These embeddings allow quantitative comparison of any two locations globally. Examples include AlphaEarth, TESSERA, and Clay, each capturing different aspects of landscape similarity depending on the satellite data and training approach used.

Geospatial Foundation Model

A large AI model pre-trained on satellite imagery and earth observation data. These models learn general representations of landscape characteristics, forest condition, and land-use patterns from billions of observations. Different models capture different aspects of landscape similarity: some emphasise spatial context, others temporal patterns such as seasonal change over multiple years.

Historical Baseline

A baseline constructed by projecting historical deforestation rates into the future. Historical baselines assume the past predicts the future, but they are not flexible enough to handle unseen confounders or rapid changes in policy, economics, or environmental conditions, which can lead to overcrediting when deforestation pressure diverges from historical trends. Risk maps are also a form of historical baseline, and whilst they can more effectively resolve the spatial challenges of reference area selection, they cannot accurately reflect unseen confounders that occur after the model training period.

ICVCM

The Integrity Council for the Voluntary Carbon Market, an independent governance body that sets the Core Carbon Principles and assesses carbon-crediting programmes against them. ICVCM-approved programmes and methodologies carry a Core Carbon Principles label intended to signal high-integrity credits.

ICVCM Core Carbon Principles

The Integrity Council for the Voluntary Carbon Market's framework defining high-integrity carbon credits across ten principles, covering governance, emissions impact, and sustainable development. Carbon-crediting programmes and methodologies that meet the Core Carbon Principles can carry the CCP label, intended as a quality signal to buyers.

IETA

The International Emissions Trading Association, the principal industry body representing companies involved in carbon markets, including project developers, traders, financial institutions, and corporate buyers. IETA advocates for emissions trading and convenes practitioners around policy and methodology development.

Improved Forest Management (IFM)

A category of carbon project that changes forest management practices, for example extended rotation, reduced impact logging, or stocking enhancement, to increase carbon storage in existing forests. Verra's IFM portfolio sits within VCS and is being expanded to include enhanced sequestration interventions under draft methodology M0274.

Jurisdictional and Nested REDD+ (JNR)

Verra's framework for REDD+ programmes operating at the jurisdictional level (state, province, or country) and the rules for nesting project-level activity inside those jurisdictional accounting frameworks. JNR aligns project crediting with broader land-use change accounting at scale, reducing double-counting risk between projects and host-country inventories.

Leakage

The risk that a carbon project displaces deforestation or degradation to areas outside the project boundary rather than preventing it entirely. If protecting one forest area shifts pressure to a nearby unprotected area, the net climate benefit may be reduced. Robust project design and monitoring aim to detect and account for potential leakage effects.

M0274

Verra's draft methodology Enhanced Forest Sequestration with Dynamic Baselines Using Randomized Control Trials, currently in public comment. M0274 broadens Verra's Improved Forest Management portfolio to credit enhanced sequestration projects whose named interventions include mycorrhizal inoculation, removal of competing vegetation (such as woody vines), and mixed-species planting. The methodology uses internal randomised control plots inside the project area, sense-checked against external reference areas in the surrounding landscape.

Nature-Based Solutions (NbS)

Actions to protect, sustainably manage, and restore natural or modified ecosystems that address societal challenges while providing human wellbeing and biodiversity benefits. Forest carbon projects are a subset of NbS, alongside wetland restoration, mangrove conservation, and other ecosystem-based approaches.

Overcrediting

When a carbon project issues more credits than the actual climate benefit achieved. Typically caused by an inflated baseline that overestimates the deforestation that would have occurred without the project, leading to credits that represent emissions reductions that never actually happened. Overcrediting is most likely to occur unintentionally when a project uses a historical baseline built on forecasts that no longer reflect current conditions.

Performance Benchmark

A VM0047 mechanism that adjusts a project's baseline based on observed outcomes in the matched reference areas during the crediting period. The benchmark updates over time using new stocking-index data from the reference areas, but the underlying matched control plots themselves are fixed at validation.

ramet5

belian.earth's counterfactual baseline service. Matches a project area to reference areas that share its biophysical and socio-economic characteristics, then uses those references to estimate what would have happened without the conservation intervention. Sits upstream of the prescribed matching procedure in methodologies such as VM0047, making reference area selection reproducible, auditable, and vendor-agnostic.

REDD+

Reducing Emissions from Deforestation and Forest Degradation, plus conservation, sustainable management of forests, and enhancement of forest carbon stocks. Originally a UN framework that creates financial value for carbon stored in forests by offering incentives to developing countries to reduce emissions from forested lands. The term is also widely applied to voluntary carbon market projects registered under Verra methodologies (such as VM0048) that reduce emissions from deforestation.

Reference Area

A geographic area used as a comparison for a carbon project site. Reference areas should share similar characteristics with the project area, such as forest type, altitude, and accessibility, with comparable pre-intervention protection levels. The quality of reference area selection directly determines the credibility of the baseline.

Reference Area Matching

The process of identifying suitable comparison areas for a carbon project. Traditional approaches rely on geographic proximity or expert judgement. More recent approaches use automated models to identify areas with genuinely similar landscape characteristics. This can produce more credible counterfactual baselines because the matched areas have not been manually chosen and therefore could not have been selected to favour a particular outcome.

Spaceborne LIDAR (GEDI)

The Global Ecosystem Dynamics Investigation, a NASA LIDAR instrument on the International Space Station that measures forest canopy height and structure from space. GEDI provides direct measurements of forest vertical structure, which correlates strongly with biomass. Unlike optical satellite imagery, LIDAR physically measures canopy height rather than inferring it from reflected light.

Stocking Index

The historical biomass or canopy-cover index used in Verra's VM0047 methodology to characterise the pre-project state of an ARR site. The stocking index defines eligibility, meaning which areas qualify as forest under the methodology, and informs the baseline against which carbon recovery is measured during the crediting period.

Synthetic Control Method

A causal inference technique that constructs a weighted combination of comparison units (from a donor pool) to create a “synthetic” version of the project area. The synthetic control estimates what the project area's deforestation trajectory would have been without intervention.

TESSERA

A Cambridge-built geospatial foundation model for earth observation. TESSERA produces embeddings of the land surface used in similarity matching and downstream analysis, including forest carbon project assessment.

VCS (Verified Carbon Standard)

Verra's voluntary carbon market standard, the largest single body of voluntary credits issued globally. The VCS programme governs registration, validation, verification, and issuance for projects across the avoided deforestation, restoration, agriculture, and energy sectors. Specific project types are credited under approved methodologies such as VM0047 (afforestation, reforestation, and revegetation) and VM0048 (REDD+).

VM0047

Verra's methodology for Afforestation, Reforestation, and Revegetation (ARR). Used for planting projects that generate carbon removal credits by establishing or restoring forest cover. Requires projects to demonstrate that carbon sequestration would not have occurred without the project intervention.

VM0048

Verra's methodology for Reducing Emissions from Deforestation and Forest Degradation. Replaces earlier REDD+ methodologies. Under VM0048, Verra establishes project baselines using jurisdictional deforestation data combined with an assessment of deforestation risk in the project area, rather than allowing project developers to set their own baselines from self-selected reference areas.