Assessment of Verra's new VM0047 ARR methodology
Here are some key takeaways
Verra has released a new methodology, VM0047, for projects in the Afforestation, Reforestation & Restoration sub-sector, with the introduction of dynamic performance benchmarks (dynamic baselines).
Other changes include a clearer definition of how leakage is quantified and the incorporation of a wider range of carbon pools in carbon accounting.
Overall, we believe this new methodology contains many improvements, while a wide range of credit efficacy remains possible given limitations in the area-based sampling approach.
This week, Verra announced the release of its new methodology for the Afforestation, Reforestation & Restoration (ARR) sub-sector, VM0047- Afforestation, Reforestation And Revegetation, which has been under development for three years. VM0047 will replace the Clean Development Mechanism methodologies AR-ACM0003 (Afforestation and reforestation of lands except wetlands) and AR-AMS0007 (Afforestation and reforestation project activities implemented on lands other than wetlands). These two methodologies are being phased out for VCS accredited projects, and Verra is requiring that new ARR projects and those under development using AR-ACM0003 or AR-AMS0007 be listed as ‘under validation’ by December 28 this year, and the validation must be completed by that date in 2024. Verra is discussing how to migrate ARR projects employing older methodologies to VM0047. Existing ARR projects may have to adopt the new methodology at the end of their current crediting period.
Verra also released a new leakage module, VMD0054 Module for Estimating Leakage from ARR Activities. This module must be employed by projects registered under VM0047, and is exclusive to this methodology. It focuses on the net production replaced in the market, the amount of new forest land cleared for displaced production outside the project area, and the associated carbon stock emissions.
We have highlighted key changes in the new methodology compared to the previous methodologies in Table 1.
Table 1. Some of the key changes in Verra’s VCM0047 from previous approved methodologies under the standards body.
Our thoughts on key changes in the methodology
While BeZero finds that a project’s choice of methodology does not necessarily influence its final BeZero Carbon Rating (BCR), we believe that the innovations and improvements made by projects in response to updated methodologies can lead to higher ratings for new projects or newer vintages of existing projects. BeZero welcomes the development of VM0047, as it leads the way in establishing greater transparency in multiple factors relating to ARR projects’ carbon efficacy, such as potential and prior land use.
What we like
By incorporating two quantification approaches - area-based or census-based - VM0047 allows for flexibility in applying the methodology to different ARR projects with different project activities, which could lead to more appropriate baselines being set for project-specific circumstances. The area-based approach paves the way for control-plot-based dynamic performance benchmarks for the wider Nature-Based Solutions sector group. These types of benchmarks are considered by many VCM stakeholders to be synonymous with ‘dynamic baselines’. This use of performance benchmarks will allow for projects to account for the natural regeneration of biomass that would have occurred in the absence of the project, which projects often overlook or assume to be zero, but which our analysis finds can be a major source of over-crediting risk for ARR projects.
BeZero believes this approach to baselines is best practice, and we already conduct our own dynamic analysis to evaluate project efficacy. For example, we analyse the national, regional and local (50-km buffer) plantation extent and tree gain at the start of, and over the course of, a project. We also look at the role of policy in plantation expansion, or the establishment of agroforestry and silvopastoral systems, depending on the project activities. These analyses both conduct a comparison of the baseline scenario against a dynamic backdrop.
VM0047 includes other changes to additionality testing, such as the introduction of a common practice benchmark under the census-based approach, under which project activities with an adoption rate of 15% or greater at a national or sub-national level (depending on relevant policy) are not considered additional. This is part of a wider movement by Verra to establish common practice benchmarks, also seen in their VM0042 methodology, released in 2020.
The explicit documentation of the census approach also allows for the easier establishment and continuation of small-scale afforestation, agroforestry, and silvopastoral projects. Similar approaches have historically been employed by projects such as the The International Small Group & Tree Planting Program projects, but have not previously been clearly documented in many ARR methodologies.
To be eligible to use the census-based approach under VM0047, a project’s activity area must have been classified as non-forest for at least 10 years prior to the project’s implementation. This is something we already evaluate as part of our assessments of additionality and over-crediting risk, and can be found in some standards bodies’ eligibility criteria, including CDM and VCS. However, this new requirement differs from the general VCS standard requirement, which only stipulates that projects cannot clear or convert native ecosystems within 10 years prior to the proposed project start date. The new stricter requirement in VM0047 ensures that projects cannot clear forest prior to the project with the goal of accessing carbon finance from replanting. For example, our analysis of a China-based afforestation project, developed under the AR-ACM0003 methodology, found considerable vegetation cover in the project area before the project began, which in our view suggests significant additionality risk for that project.
With the release of an accompanying leakage module, it is likely that projects employing VM0047 will provide more transparency regarding prior land use. This module focuses on lost productivity from prior agricultural land use based on a historical period of three years or one crop rotation (whichever is greater). While this has been accounted for in other methodologies historically, neither AR-ACM0003 or AR-AMS0007 stipulate a specific timespan. Furthermore, many projects do not provide the depth of analysis stipulated by this methodology, and therefore are subject to information risk. Under VM0047, leakage effects are considered for five years after the project start date in order to capture the period in which the effects of leakage from displaced agriculture are expected to occur. However, the methodology does not acknowledge positive leakage, which we commonly observe amongst ARR projects such as Timor Leste-based afforestation project, where tree planting could lead to slope stabilisation, enhancing soil carbon stocks outside of the project area.
Stipulating the inclusion of the non-woody biomass and soil carbon pools (under the circumstances highlighted in Table 1), will likely enhance the accuracy of estimated carbon removals in projects’ baseline scenarios. Furthermore, it will allow for better monitoring and reporting of disturbances during the planting and management periods. Non-woody biomass and soil carbon pools are already accounted for by some methodologies such as the Woodland Carbon Code (WCC). However, our rating of one UK-based afforestation project accredited under the WCC, found that the methodology likely did not accurately account for the presence of carbon-rich peaty soils, which contributed to our view that the project faces significant risk of over-crediting. Both of these factors are risks that we explore when assessing over-crediting.
Room for improvement
BeZero Carbon welcomes the introduction of VM0047, especially for its dynamic performance benchmarks and improved leakage assessment. However, we note some areas for improvement for dynamic performance benchmarks. At present, there appears to be no handling of uncertainty in the matching process, and the matching of project areas to controls involves only one factor - the stocking index - albeit at three time points prior to the project’s start. Thus, it is possible that the project area and controls may differ significantly in environmental and geospatial conditions that are not considered by VM0047. As a result, there is a possibility of the matching quality being compromised. We believe that Verra could enhance transparency by requiring projects using VM0047 to disclose the location of control plots, to allow for independent assessment of the dynamic performance benchmark.
Our own work on dynamic baselines incorporates a rigorous treatment of uncertainty in the statistical matching process. In addition, our Geospatial and Data Science teams collaborate closely with ratings scientists, to ensure that each project-level calibration of our dynamic baseline incorporates the latest project-specific information. By matching on multiple correlates of change, such as local climate and soil conditions, distance to roads, population pressure, and governance, we integrate project-level nuance with satellite monitoring, machine learning, and statistical handling of uncertainty.
We welcome the improvements in quality that are brought about when standards bodies - Verra, in this case - identify the risks inherent in older methodologies and develop new ones to address these concerns. Many of the changes introduced by VM0047 could improve the carbon accounting and carbon efficacy of ARR projects. In particular, we find that the method’s major improvements are the introduction of dynamic baselines, a definitive common practice benchmark, and clearer leakage monitoring guidelines. Nevertheless, we believe there are still improvements that could be made. Independent risk assessment of individual projects and vintages remains key to determining the quality of issued credits.