# Forest Inventory Plots for ARR Carbon Projects: How Many, How Big, How Often

Last updated: 2026-11-01

> Work out how many inventory plots an ARR carbon project needs for ±10% at 90% confidence, or ±20% for Gold Standard vintages to 2025, how big and how often.

Date: November 1, 2026

> [ARR Carbon Guide](https://restorationscope.com/blog/category/carbon): take your ARR project from standard to verification, step by step.

An ARR carbon project needs enough forest inventory plots to estimate biomass to its standard's precision. Verra VM0047 deducts uncertainty above 10% at 90% confidence on the change in carbon stocks, and the CDM sample-size tool defaults to ±10% at 90%. Gold Standard's A/R methodology v2.1 asked ±20% at 90% per modelling unit; from 2026 vintages, its STARR draft (not yet final) would set ±10%. The number follows from how much plots vary: n = (t × CV ÷ E)². If plots vary by 40% of the mean, ±10% at 90% needs 44 plots, or 49 with 10% spare.

This guide is for the MRV or forestry lead planning an inventory before validation. It covers how many plots, how big, permanent or temporary, how often, and what a missed target costs. Restoration Scope's free [forest inventory plot calculator](https://restorationscope.com/tools/forest-inventory-plot-calculator) runs the same numbers for your project, in your browser.

**Checked 1 November 2026** against [CDM AR-TOOL03 v02.1.0](https://cdm.unfccc.int/methodologies/ARmethodologies/tools/ar-am-tool-03-v2.1.0.pdf), [Verra VM0047 v1.1](https://verra.org/documents/vm0047-afforestation-reforestation-and-revegetation-v1-1/), Gold Standard's [A/R GHG methodology v2.1](https://globalgoals.goldstandard.org/standards/403_V2.1_LUF_AR-Methodology-GHGs-emission-reduction-and-Sequestration-Methodology.pdf) and the [IPCC Good Practice Guidance for LULUCF](https://www.ipcc-nggip.iges.or.jp/public/gpglulucf/gpglulucf_files/Chp4/Chp4_3_Projects.pdf), section 4.3. A planning guide in our own words; your methodology and your validation and verification body (VVB) decide what is enough.

## How many plots do you need?

The CDM tool for sample plots in A/R projects (AR-TOOL03) works the count out from three numbers:

- **t**, the Student's t-value for your confidence level: 1.645 at 90% and 1.96 at 95% to start with, larger when the answer is under 30 plots.
- **CV**, how much plot biomass varies: the standard deviation between plots divided by their mean, as a percentage.
- **E**, the error you can accept, as a percentage of the mean: 10 for ±10%.

For one stratum, when the plots cover under 5% of the area, the tool's simplified equation (equation 2) is:

$$
n = \left(\frac{t \cdot \mathit{CV}}{E}\right)^{2}
$$

(AR-TOOL03, eq. 2)

With a CV of 40% and a target of ±10% at 90%: $\left(1.645 \times 40 / 10\right)^{2} = 43.3$. The full equation, which also counts how many plots fit in a 500 ha project of 0.05 ha plots, gives 43.1, rounded up to 44 plots.

Two corrections matter for small samples and small sites. When the answer is under 30 plots, the tool works it out again with the t-value for that many plots minus one, which adds a few. When plots cover more than about 5% of the area, equation 1's finite-population term takes a few away. Above 5%, AR-TOOL03 then applies a further cut (equation 3). Equation 1 already holds a finite-population term, so our calculator leaves equation 3 out: on small sites it can give many more plots than the tool.

Try it with your own numbers:

[Forest inventory plot calculator](https://restorationscope.com/tools/forest-inventory-plot-calculator): enter how much your pilot plots vary, the area, the plot size and the precision target to get the plot count.

Plots needed for a 500 ha project with 0.05 ha plots, before spares:

| How much plots vary (CV) | ±10% at 90% | ±20% at 90% | ±10% at 95% |
| --- | --- | --- | --- |
| 20% | 14 | 9 | 19 |
| 30% | 27 | 9 | 35 |
| 40% | 44 | 14 | 62 |
| 60% | 97 | 27 | 137 |
| 80% | 171 | 44 | 240 |

Under 30 plots, the t-value comes from the first count rounded up, minus one. At low variation that t-value adds most of the plots, which is why a CV of 20% and 30% both need 9 at ±20%.

Once the plots cover a small share of the project, the count barely depends on its size. With the same CV and plot size, 50 ha needs 42 plots and 500 ha needs 44. What drives it is how much the plots vary and how tight the target is. Above 30 plots, halving the error roughly quadruples them: at a CV of 40%, ±10% needs 44 plots and ±5% needs 171.

## What precision does each standard ask for?

| Standard | Precision target | Above the target |
| --- | --- | --- |
| Verra VM0047 v1.1 | 10% at 90% confidence, on the change in carbon stocks | Each point of uncertainty above 10% adds a point to the deduction; no credits above 100% |
| CDM AR-TOOL03 v02.1.0 | ±10% of the mean at 90% by default, unless the methodology says otherwise | Set by the methodology that uses it |
| Gold Standard A/R GHG methodology v2.1 (credits up to 2025 vintage) | ±20% at 90% for each modelling unit | Error above 20% is deducted from the mean |
| Gold Standard STARR (draft, consulted April–May 2026) | ±10% at 90%, on the change in biomass for large projects with sample plots | Total uncertainty above 10% is deducted |

Under the STARR draft, default allometric equations add a 10% model error, so ±10% sampling alone still costs about 4%.

Gold Standard is moving to Paris-aligned methodologies: credits from 2026 vintages onward need one, and 30 June 2026 was the cutoff for submitting a new project for preliminary review under an older methodology. Projects submitted by then can use the older methodology for design certification and credits up to 2025 vintage. For 2026 vintages onward, existing projects must also pass a Paris-alignment design change and use the new methodology; issuance waits until it is published and applied, then catches up. Its draft ARR methodology, STARR, would tighten the target to ±10% once final, so an inventory planned now for 2026 or later removals should aim for ±10%.

For improved forest management, [ACR's non-federal US methodology](https://acrcarbon.org/wp-content/uploads/2022/07/ACR-Methodology-IFM-on-Non-Federal-US-Forestlands-v2_1-20240701.pdf) deducts total uncertainty, baseline and project together, above ±10% at 90%, and the [Climate Action Reserve's US protocol](https://climateactionreserve.org/wp-content/uploads/2023/07/Final_Forest_Protocol_V5.1_7.14.2023.pdf) deducts sampling error above ±5% at 90% for projects that aren't aggregated. [Verra VM0045](https://verra.org/documents/vm0045-improved-forest-management-using-dynamic-matched-baselines-from-national-forest-inventories-v1-2/) deducts above 15% at 95%, on the estimated removals against its matched baseline. The [IFM standards comparison](https://restorationscope.com/blog/ifm-carbon-standards) has those rules.

Sources: [AR-TOOL03 v02.1.0](https://cdm.unfccc.int/methodologies/ARmethodologies/tools/ar-am-tool-03-v2.1.0.pdf), equations 1–4; [VM0047 v1.1](https://verra.org/documents/vm0047-afforestation-reforestation-and-revegetation-v1-1/), section 8.5, equation 28; [Gold Standard A/R GHG methodology v2.1](https://globalgoals.goldstandard.org/standards/403_V2.1_LUF_AR-Methodology-GHGs-emission-reduction-and-Sequestration-Methodology.pdf), section 3.11.5; [Gold Standard Paris Agreement alignment](https://globalgoals.goldstandard.org/gs4gg-paris-agreement-alignment/); [STARR consultation draft](https://goldstandard.cdn.prismic.io/goldstandard/aeiFSsBOoF08xMq4_GSARR-Meth_V1PC.pdf), sections 14 and 15.

## Why does VM0047's 10% apply to the change?

VM0047 doesn't fix a number of plots. It measures uncertainty on the carbon gained since the project started, ΔC, combining the standard errors of the initial (t = 0) inventory and the current one. On land that started with little woody biomass, and so little uncertainty at t = 0, that is close to the precision of today's stock, so the calculator is a fair first estimate.

On land that started with more biomass, the same plots give a wider error on the change. Permanent plots help when a plot's two measurements move together, as they usually do: VM0047's formula subtracts that correlation, which only permanent plots have.

For the ex-ante estimate at validation, VM0047 deducts at least 10% for uncertainty whatever the plots show. The [VM0047 field data guide](https://restorationscope.com/blog/vm0047-arr-field-data) covers what else the methodology asks you to record.

## Where does the CV come from?

From a pilot. IPCC good practice suggests about 10–15 plots in each stratum may be enough, measured in stands like the ones you will monitor, then dividing the standard deviation of plot biomass by its mean. AR-TOOL03 also accepts existing inventory data from the area or a similar one.

Measure the pilot in a stand like the one you will monitor, at the age your plots will reach: IPCC's example is 10–15 plots in an existing 20-year-old forest for a 20-year project, with the same species mix and spacing. If the pilot plots vary more than you expected, that is the number to plan with.

## How do strata reduce the plot count?

Strata group land that grows alike: one planting year, species mix, soil or site. Variation within each stratum is smaller than across the whole project, so well-chosen strata need fewer plots for a project-wide target. Requiring the target in every modelling unit, as Gold Standard's v2.1 does, can need more.

AR-TOOL03 then allocates plots to each stratum in proportion to its area times its standard deviation (equation 4): a large, variable stratum gets the most plots. Our calculator gives each stratum at least two plots, the fewest that can estimate its variance. Gold Standard's ±20% holds for each modelling unit, so with that target the calculator treats each stratum as a modelling unit and works it out on its own. For trees planted over several years, IPCC good practice treats each cohort of similar age as its own population, with no more than two or three age classes in one cohort.

## How big should each plot be?

For even-sized stands, IPCC good practice gives 100 m² for dense stands of 1,000 trees per hectare or more, up to 600 m² for sparse ones. Where tree sizes vary a lot, nested plots measure small trees in an inner circle and large trees across the whole plot. The minimum diameter at breast height is often 5 cm; young plantings are mostly smaller trees, which IPCC counts in a subplot.

Circular plots are quick to lay out from one centre point:

| Plot area | Circle radius |
| --- | --- |
| 100 m² (0.01 ha) | 5.64 m |
| 500 m² (0.05 ha) | 12.62 m |
| 1,000 m² (0.1 ha) | 17.84 m |

IPCC good practice ties plot size to how much plots vary, and bigger plots usually vary less, so plot size and CV move together. If your pilot used one size, estimate the CV for the size you plan to use before you fix the count.

## Permanent or temporary plots?

Permanent plots, measured again at each monitoring event, usually estimate change more efficiently than new plots each time, because each plot is compared with itself. Gold Standard's [STARR draft](https://goldstandard.cdn.prismic.io/goldstandard/aeiFSsBOoF08xMq4_GSARR-Meth_V1PC.pdf) asks large-scale projects for stratified permanent sample plots (Method A); smaller projects are directed to count every tree (Method B), though its Table 14 also allows Method A.

Permanent plots only work if a crew can find them years later:

1. **Lay them out on a grid with a random start.** A systematic grid spreads plots across the project; the random start keeps the sample unbiased.
2. **Install 10% spare.** IPCC guidance recommends installing 10% more than the calculated count, for plots that can't be found again.
3. **Mark the centre and record it.** A buried marker or painted post, a GPS point and a photo. Don't move a centre because the spot is inconvenient.
4. **Tag the trees.** A numbered tag on each tree lets the next crew measure the same stem.

## How often are plots remeasured?

- **Verra VM0047:** biomass plots every five years or more often; stocking-index values at least once a year for area-based projects.
- **Gold Standard A/R v2.1:** the inventory is repeated at least before every Performance Certification.
- **Quality checks:** IPCC good practice has an independent crew remeasure one plot in every 8–10, and checks 10–20% of plots at the end.

Sources: [IPCC GPG-LULUCF 2003, sections 4.3.3.4 and 4.3.4.2](https://www.ipcc-nggip.iges.or.jp/public/gpglulucf/gpglulucf_files/Chp4/Chp4_3_Projects.pdf); [VM0047 v1.1](https://verra.org/documents/vm0047-afforestation-reforestation-and-revegetation-v1-1/), monitoring parameters; [Gold Standard A/R GHG methodology v2.1](https://globalgoals.goldstandard.org/standards/403_V2.1_LUF_AR-Methodology-GHGs-emission-reduction-and-Sequestration-Methodology.pdf), section 3.11.4.

## What does missing the target cost?

Credits, not the project. Under VM0047, an uncertainty of 14% on the change gives a 4% uncertainty deduction. It applies to cumulative removals, so its effect on one monitoring period's issuance follows the methodology's equation 32 and can be larger than 4%. Under Gold Standard's A/R methodology v2.1, a mean of 100 m³/ha estimated at ±23% is reported as 97 m³/ha.

So the plot count is a trade-off you can price: each extra plot costs a crew's time, and each point of uncertainty above the target costs credits. The calculator's "If one thing changes" table shows how fast the count moves.

## Where software fits

A plot design is only as good as the records behind it: which plot is where, which trees were measured in it, by whom and when, and which plots a crew couldn't find. Kept in spreadsheets and phone galleries, those records drift apart between monitoring events.

Restoration Scope gives field teams offline capture of tree and plot records with GPS and photos, keeps the history of inspections and field updates on each record, and exports tables as CSV for your VVB. It doesn't design your sample or calculate credits. See [Restoration Scope for carbon projects](https://restorationscope.com/solutions/carbon).

**A monitoring plot on a phone.**

The plot on the map with each tree recorded inside it, and the plot's record below. Demo project: names and numbers are made up.

![Demo monitoring plot MP0009 on a phone: the plot circle on the satellite map with five tree points inside it, above the plot record with trees monitored and species counts](https://restorationscope.com/images/landing/mobile-geospatial/demo-plot-point.png)

More guides: [getting VM0047 records ready for verification](https://restorationscope.com/blog/vm0047-verification-readiness), [ARR carbon standards compared](https://restorationscope.com/blog/arr-carbon-standards) and [ARR carbon monitoring software](https://restorationscope.com/blog/arr-carbon-monitoring-software).

**Disclosure:** Restoration Scope makes field and monitoring software for carbon projects. We are not a validation or verification body (VVB), and we don't issue credits.

## Questions about forest inventory plots

### How many inventory plots does a reforestation project need?

It depends on how much biomass varies between plots and the precision your standard asks for. The CDM sample-size tool (AR-TOOL03) gives n = (t × CV ÷ E)² for one stratum: with plots that vary by 40% of the mean, ±10% at 90% confidence needs 44 plots, and ±20% needs 14 once the tool's small-sample t-value is used. IPCC guidance recommends 10% spare plots on top.

### Does VM0047 set a minimum number of plots?

No. Under its area-based approach, VM0047 v1.1 asks for plot-based sampling of woody biomass and deducts uncertainty above 10% of the carbon stock change at 90% confidence. The 30-plot minimum in VM0047 is for the remote-sensing plots of its performance benchmark, not the field biomass plots.

### What size should forest inventory plots be?

IPCC good practice gives 100 m² for dense stands of 1,000 trees per hectare or more, up to 600 m² for sparse ones in even-sized stands, with nested plots when tree sizes vary a lot. A 0.05 ha (500 m²) circular plot has a radius of 12.62 m.

### How often do ARR inventory plots have to be remeasured?

Under Verra VM0047, biomass plots are measured every five years or more often. Under Gold Standard's A/R methodology v2.1, the inventory is repeated at least before every Performance Certification.

### Should inventory plots be permanent or temporary?

Permanent plots, measured again each time, estimate the change in carbon with fewer plots, because each plot is compared with itself. VM0047's uncertainty formula gives permanent plots credit for that correlation, when a plot's measurements move together.

### What happens if the inventory misses the precision target?

Excess sampling uncertainty costs credits; the inventory's other requirements still apply. Under VM0047, each point of uncertainty above 10% adds a point to the deduction on cumulative removals, so one period's issuance can drop by more, and no credits are issued above 100%. Under Gold Standard's A/R methodology v2.1, error above ±20% is deducted from the mean; its STARR draft would deduct total uncertainty above 10%.