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.
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.
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.
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.
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.
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.
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.
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.
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.
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.