Insights · Methodologies
VM0047 explained: the ARR methodology in plain English
VM0047 is the rulebook Verra uses for tree-planting carbon projects: afforestation, reforestation and revegetation, including agroforestry. It decides what land qualifies, how removals are measured and what is subtracted before credits are issued. This is what it says, without the equations, and what it means for your project.
What VM0047 is
VM0047 is a methodology under Verra's Verified Carbon Standard (VCS). A methodology is the detailed recipe a project must follow to turn growing trees into carbon credits: which land is in, which carbon pools count, how to measure them, what the baseline is and what gets deducted.
Version 1.0 was published in September 2023, developed by TerraCarbon and Silvestrum Climate Associates. Version 1.1, developed by Verra with support from TerraCarbon, has applied since 14 May 2025. It builds on the UN Clean Development Mechanism's large-scale tree-planting methodology, AR-ACM0003. VCS projects still using the older CDM methodologies had to complete validation by 30 June 2025, so for new tree-planting projects under the VCS, VM0047 is the methodology.
Projects already designed under version 1.0 have a grace period, set out on Verra's methodology page: pipeline listing as “under validation” had to be requested by 31 December 2025, and registration must be requested by 31 December 2026. Everything else follows version 1.1, which also added the 50-trees-per-hectare limit for the census-based approach and the option to use GPS points or physical markers to identify each tree. In October 2026 Verra also published guidance for existing projects moving to VM0047 from other methodologies, such as AR-ACM0003; if you are transitioning a registered project, read that first.
- StandardVerra VCS
- Current versionv1.1, since 14 May 2025
- ActivityARR, incl. agroforestry
What it covers
VM0047 applies to activities that establish, increase or restore vegetative cover. The trees can be planted, sown, or brought back through assisted natural regeneration, such as fencing out grazing animals or cutting lianas, as long as vegetative cover measurably increases.
There are two kinds of eligible land:
- Non-forest land. Grassland, cropland, pasture, degraded land. This is classic afforestation and reforestation, and it is what our screenings study.
- Existing forest that has not been managed for wood products in the past ten years, where the project adds carbon. This is only possible with the area-based approach described below.
Revegetation means woody cover that does not reach the forest definition: shrubs, or sparse trees on dry land. Agroforestry is also in scope; scattered trees on farms typically use the census-based approach. Projects on organic soils or wetlands use VM0047 for the trees and a wetland methodology for the other carbon pools.
Some things are excluded. Under the area-based approach, land that met the definition of managed forest at any point in the ten years before the start date is out, and so is clearing existing woody vegetation by harvesting timber or by degrading native ecosystems. For the full eligibility picture, see reforestation (ARR) feasibility and the forest definition and the 10-year rule.
Area-based vs census-based: the two VM0047 approaches
The most important design choice in VM0047 is the quantification approach. The two are mutually exclusive for any piece of land, and the choice made at the start date holds for the whole crediting period. A project can use both on separate areas, kept at least 10 metres apart.
| Area-based | Census-based | |
|---|---|---|
| Typical project | Block planting or regeneration that changes land cover | Scattered trees on farms, along boundaries, in towns |
| Planting density | Not capped; projects that could use census-based must do so | At most 50 planting units per hectare |
| Land use | Changes (e.g. pasture becomes forest) | Continues (e.g. cropping goes on under the trees) |
| How carbon is measured | Field plots, scaled up by area | Every tree counted; a sample measured, scaled up by the count |
| Baseline | Dynamic performance benchmark from control plots | Zero |
| Leakage | VMD0054; may not be assumed negligible | Exempt, because land use does not change |
The census-based approach has extra conditions. Only direct planting counts. The trees must be on land with less than 10% pre-existing woody cover, or on continuously cropped land, settlements or “other land” in IPCC terms. Every tree needs a GPS point or a physical marker, and a tree that cannot be found at monitoring is counted as dead. Agroforestry projects often fit here; read more in agroforestry carbon credits.
The dynamic performance benchmark and control plots
The area-based approach does not assume that nothing would have grown without the project. Instead it watches what actually grows on similar land nearby. This is the dynamic performance benchmark, and it is the most talked-about part of VM0047.
- Project plots. The project area is divided into equal units of 0.01 to 10 hectares, and at least 30 are sampled at random.
- Control plots. For each project plot, matched control plots are picked outside the project, within a donor pool that shares the same jurisdiction, ecoregion, policy environment and land tenure types.
- Stocking index. A remote sensing metric that correlates with above-ground carbon, such as a vegetation index from Landsat or canopy height from lidar, is tracked on both sets of plots.
- Comparison. The benchmark is the ratio of the change in stocking index on control plots to the change on project plots. Growth that also happens on control plots is not credited.
Two details matter. Remote sensing is used only for the relative change between project and control plots; the carbon itself is measured in the field. And the benchmark is updated at each verification, or every five years, whichever comes first. That is why it is called dynamic. If the land around the project starts regrowing on its own, the credited share shrinks. For a longer walk-through, see the dynamic baseline in VM0047.
Additionality and leakage
Credits must come from removals that would not have happened anyway. VM0047 checks this differently per approach. Area-based projects show regulatory surplus (no law already requires the planting), pass the performance benchmark test, and run an investment analysis unless the project earns nothing other than carbon credits. Census-based projects show regulatory surplus, an investment analysis and common practice. More in additionality for tree-planting projects.
Leakage is the farming or grazing that moves elsewhere when trees take over the land. Area-based projects must estimate it with the module VMD0054. It can be large. In our anonymised sample study of a dryland project in East Africa, leakage under the VMD0054 default values for displaced land use came to 55,310 tCO2e, 79% of the 69,779 tCO2e gross estimate over 40 years. That is a screening estimate, not a validated figure, but it shows why leakage belongs in the first conversation, not the last.
What VM0047 means for developers
- Measure before you clear. The start date is the earlier of site preparation or the land-use change (for example, stopping grazing). Where site preparation starts the project, the starting carbon stock must be measured no more than two years before it. Plots that were never measured cannot be reconstructed later, except pre-existing woody biomass by remote sensing.
- Choose the approach early. It cannot be switched later for the same land.
- The forecast must be conservative. The project description needs an ex-ante estimate for the whole crediting period, from growth and yield models that conservatively represent the project, with an uncertainty deduction of at least 10%, and, for area-based projects, a forecast benchmark, which may extend the control plots' trend over the ten years before the start.
- Budget for remote sensing and field work. The benchmark needs a repeatable, documented control-plot design. Removals need field plots at each verification.
None of this is visible in a hectares-times-growth-rate spreadsheet. It is why the gross and net figures of the same project can be far apart.
What VM0047 means for a feasibility screening
A carbon feasibility study can test most of what VM0047 asks before a project spends on design, and must be honest about the rest. Ours follows VM0047; it is a desktop screening, not a validation.
What a screening can do: check the land against the eligibility rules with satellite data, estimate gross removals with IPCC 2019 Refinement Tier 1 growth rates in conservative, average and optimistic cases, apply VMD0054 default values for leakage, and show a buffer and an uncertainty deduction as planning assumptions. Our screening applies a 10% uncertainty deduction, the minimum VM0047 sets for ex-ante estimates, and uses 15% for the buffer, a planning figure taken from the median of 198 registered ARR projects in our registry data.
Frequently asked questions
VM0047 is Verra's methodology for afforestation, reforestation and revegetation (ARR) under the Verified Carbon Standard. It sets the rules for measuring the carbon dioxide that planted or regenerating woody vegetation removes, how to set the baseline, how to show additionality and how to handle leakage. Version 1.0 was published in September 2023 and version 1.1 has applied since 14 May 2025.
The area-based approach is for projects that turn non-forest land into forest, or add carbon to existing forest that is not managed for wood. It samples plots and compares the project with control plots seen by satellite. The census-based approach is for scattered planting of at most 50 trees per hectare on land whose use does not change, such as farms. Every tree is counted and tracked.
Yes, in two ways. Trees planted at low density among crops, with farming continuing, fit the census-based approach, which has no leakage deduction and a baseline of zero. Denser planting that changes the land cover is quantified with the area-based approach instead. Which one applies depends on planting density and whether the existing land use continues, and the choice is fixed for the crediting period.
No. The benchmark compares the change in a satellite stocking index on project plots with matched control plots, observed during monitoring. That data does not exist before the project starts. A desktop screening can only flag where natural regrowth is likely, which is a risk to the net figure, and say so plainly in the report.
Sources
- Verra: VM0047 Afforestation, Reforestation and Revegetation (methodology page, v1.0 and v1.1)
- Verra: VM0047 Afforestation, Reforestation and Revegetation, v1.1 (methodology document)
- Verra: VCS Standard, v4.7 (16 April 2024)
- IPCC (2019), 2019 Refinement to the 2006 IPCC Guidelines, Volume 4: Agriculture, Forestry and Other Land Use