REDD+ Uncertainty Training

Foundational Training on Uncertainty Statistics in Carbon Accounting for Jurisdictional ART-TREES Programs (V2.0)

Preface

Winrock International commissioned this training resource under its JTAP portfolio (20065-2025-ICA-03) as a Foundation Module on uncertainty quantification in jurisdictional REDD+ carbon accounting. The curriculum gives methodological guidance on quantifying, reporting and reducing uncertainty in REDD+ carbon accounting under the ART-TREES Standard (V2.0) and the IPCC 2019 guidelines (ART, 2021; IPCC, 2019).

Estimation uncertainty is treated here as a measure that shows where the largest errors lie and where credit issuance can be improved, rather than as a penalty to be managed (Camara et al., 2024; Duncanson et al., 2021; Simoes et al., 2021). A program that invests in reducing its largest sources of uncertainty earns more from crediting because the deduction shrinks, and it can also earn a higher price through greater credibility and lower verification and monitoring costs (Köhler & Huth, 2010).

Uncertainty Potential

Uncertainty reporting across REDD+ programs is incomplete, in that Butler et al. (2024) found that 91 per cent of participating countries reported activity data uncertainty but only 4 to 14 per cent reported emission factor and allometric uncertainty, which leaves most jurisdictions room to complete their methods and improve credit issuance. When emission factors and modelling assumptions are included, the uncertainty of reference levels ranges from 4.2 per cent to 262.2 per cent (Pelletier et al., 2013).

Complete uncertainty reporting serves more than compliance, because a jurisdiction that quantifies and reports every source has a baseline against which improvement can be measured and a route to higher credit revenue. Butler et al. (2024) note that a complete assessment can at first widen the confidence interval, but that the same transparency shows where method can be improved and so reduces the true uncertainty over time, which raises credit issuance and revenue. Results-based payment programs have built in mechanisms that reward complete reporting, and the FCPF Carbon Fund caps its uncertainty deduction so that no program is penalised for reporting in full. The FCPF Conservativeness Factors set the deduction by uncertainty band.¹

• No deduction if uncertainty is 15 per cent or less
• 4 per cent deduction if uncertainty is 15 to 30 per cent
• 8 per cent deduction if uncertainty is 30 to 60 per cent
• 12 per cent deduction if uncertainty is 60 to 100 per cent
• 15 per cent maximum deduction if uncertainty is above 100 per cent

The cap means that a program with very high initial uncertainty faces the same 15 per cent maximum deduction as one just above 100 per cent, so full reporting carries no extra penalty while the incentive to improve remains. The ART-TREES Standard goes further and lets a participant recover credits deducted in excess when cumulative uncertainty falls across a multi-year crediting period (Section 8, V2.0), so that both schemes reward measured improvement and investment in method.

Uncertainty Compliance

Section 8 of the ART Standard V2.0 sets six criteria (ART, 2021, p. 45).

1. Monte Carlo simulation with at least 10,000 iterations for uncertainty propagation
2. A 90 per cent confidence interval, whose half-width gives the adjustment factor
3. Conservative bias, in that systematic underestimation is acceptable and overestimation is prohibited
4. Whole-chain integration, combining activity data and emission factor uncertainties
5. Crediting period aggregation, with the option to sum uncertainty deductions across years
6. Allometry exemption, in that allometric modelling uncertainty is not mandatory and is excluded

Equation 10 gives the uncertainty deduction.

\[ UNC_t = (GHGER_t + GHGREMV_t) \times UA_t \]

Equation 11 gives the uncertainty adjustment factor.

\[ UA_t = 0.524417 \times \frac{HW_{90\%}}{1.645006} \]

Training Curriculum

The curriculum covers the four main sources of uncertainty in REDD+ carbon accounting, which are allometric equations, emission factors, activity data, and the propagation of their combined error. Winrock International works with partner jurisdictions on targeted improvements using Monte Carlo simulation tailored to each source.² Primers on uncertainty statistics and on Monte Carlo methods are given in the appendix (Uncertainty Primers), and the four chapters draw on examples and exercises from this ongoing work.

• Chapter 1, Allometric Uncertainty. Interventions in allometric estimation that reduce uncertainty by 30 to 50 per cent are presented, including local Tier 2 equations built from destructive sampling, model ensembles using Bayesian averaging, wood density and height data from LiDAR or field measurement, and standard measurement protocols with quality assurance and quality control.
• Chapter 2, Emission Factor Uncertainty. Several methods and data choices lower emission factor uncertainty, including field validation campaigns, laboratory measurement of organic carbon in soil profiles, seasonal modelling of fuel loads, moisture content and disturbance severity, and analysis of stand regeneration and updated growth curves. The choice of IPCC default emission factor for each land conversion is an often overlooked resource, and the chapter presents typical defaults ranked by their uncertainty.
• Chapter 3, Activity Data Uncertainty. Large reductions come from better reference data and data cube processing that improve image classification, from multi-temporal validation datasets that separate signal from noise, and from systematic accuracy assessment protocols.
• Chapter 4, Monte Carlo Aggregation. A sound Monte Carlo framework includes a sensitivity analysis that identifies each source of uncertainty and its share of the credit deduction. Winrock provides technical assistance with these methods, drawing on use cases to help jurisdictions choose cost-effective improvements and quality assurance suited to their forest conditions and monitoring infrastructure.

ART. (2021). The REDD+ environmental excellence standard (2.0 ed.). https://www.artredd.org/wp-content/uploads/2021/12/TREES-2.0-August-2021-Clean.pdf

Butler, B. J., Sass, E. M., Gamarra, J. G., Campbell, J. L., Wayson, C., Olguıń, M., Carrillo, O., & Yanai, R. D. (2024). Uncertainty in REDD+ carbon accounting: A survey of experts involved in REDD+ reporting. Carbon Balance and Management, 19(1), 22.

Camara, G., Simoes, R., Souza, F., Menino, F., Pelletier, C., Andrade, P. R., Ferreira, K., & Queiroz, G. (2024). Uncertainty and active learning analysis. In Satellite time series on earth observation data cubes. https://e-sensing.github.io/sitsbook/uncertainty-and-active-learning.html#uncertainty-and-active-learning

Duncanson, L., Disney, M., Armston, J., Nickeson, J., Minor, D., & Camacho, F. (2021). Aboveground woody biomass product validation good practices protocol. https://doi.org/10.5067/DOC/CEOSWGCV/LPV/AGB.001

IPCC. (2019). 2019 refinement to the 2006 IPCC guidelines for national greenhouse gas inventories (Agriculture, Forestry and Other Land Use, Vol. 4). Intergovernmental Panel on Climate Change. https://www.ipcc-nggip.iges.or.jp/public/2019rf/vol4.html

Köhler, P., & Huth, A. (2010). Towards ground-truthing of spaceborne estimates of above-ground life biomass and leaf area index in tropical rain forests. Biogeosciences (Online), 7(8), 2531–2543.

Pelletier, J., Martin, D., & Potvin, C. (2013). REDD+ emissions estimation and reporting: Dealing with uncertainty. Environmental Research Letters, 8(3), 034009.

Simoes, R., Camara, G., Queiroz, G., Souza, F., Andrade, P. R., Santos, L., Carvalho, A., & Ferreira, K. (2021). Satellite image time series analysis for big earth observation data. Remote Sensing, 13(13), 2428.

---

1. https://www.forestcarbonpartnership.org/sites/default/files/documents/fcpf_buffer_guidelines_v4.2_clean_cf28.pdf

2. See https://seamusmurphy.shinyapps.io/winrock-monte-carlo/