14 March 2023
The critical factor driving over-crediting risk in cookstove projects
Carbon Rating Scientist
Dr Kirti Ramesh
Director of Carbon Ratings
A key parameter in our assessment of over-crediting risk in Household Devices projects such as cookstoves is the fraction of non-renewable biomass (fNRB), which represents the proportion of woody biomass that is harvested unsustainably.
Previous methods developed by the Clean Development Mechanism (CDM) to estimate fNRB on a national scale may have been based on inaccurate assumptions about fuelwood harvesting.
The global average fNRB reported by projects rated by BeZero is 86%, whereas more conservative global average estimates range from 20-30%. This poses significant over-crediting risk.
Some types of carbon projects, such as improved cookstove projects, aim to avoid emissions by reducing the consumption of fuels such as woody biomass or charcoal. However, net reductions are only realised when the biomass saved is non-renewable.
If an area is harvested at a rate that exceeds its annual growth rate, this is considered non-renewable as it would lead to a net decline in biomass. If instead an area is harvested in a sustainable manner that does not exceed the annual growth rate, this is viewed as renewable. Emissions caused by burning woody biomass from renewable sources could therefore be offset by re-growth. This means there would be no overall change to carbon stocks, and no climate impact.
To determine emission reductions, projects must calculate the fraction of non-renewable biomass (fNRB): the proportion of woody biomass that would otherwise be depleted without the project activity. Past approaches for calculating emission reductions primarily relied on country-specific default values (or a similar tool for project-specific assessments) estimated by the Clean Development Mechanism (CDM) and United Nations Framework Convention on Climate Change (UNFCCC), and approved by a local designated national authority.
However, calculating fNRB is complicated, and can often lead to risks of over-crediting. This has led to new approaches to calculating fNRB, namely a choice between the CDM’s updated default value and a tool to calculate fNRB, both introduced in 2017.
Projects aiming to reduce household biomass consumption include improved cookstoves, water purification and biodigesters. Cookstove projects distribute stoves that are cleaner or more efficient than traditional technologies, so that people require less biomass for cooking. Water projects aim to reduce the amount of biomass used to purify water through boiling, using a variety of activities such as water filtration systems and rehabilitated boreholes. Biodigester projects use treated manure to generate renewable biogas, which people can burn in place of biomass.
For all such projects, it is crucial to work out the fNRB of biomass saved in order to calculate emission reductions. However, we find that understanding of fNRB tends to be uneven, with one analysis suggesting that it is the largest source of uncertainty in emission reduction calculations.¹
If projects use a non-conservative fNRB value, this can present over-crediting risks. For example, one peer-reviewed study of 51 countries investigating pan-tropical woodfuel sustainability found that in 98% of cases, the emission reductions expected by projects were overestimated due to unrepresentative fNRB assumptions.² Because of this, as well as other drivers, we often find significant over-crediting risk in most of the Household Device projects we have rated to date (Figure 2).
We find that fNRB values reported by projects in this Sector Group range from 67% (Vietnam) to 99% (Ghana) (Figure 3).
Sixteen of the 24 projects rated by BeZero relied on default country-level values approved by the CDM and UNFCCC.³ There are two key reasons why these originally assessed default values may, in our view, not be appropriate.
National versus local assumptions
Country-level defaults do not accurately reflect project-specific locations.
Default fNRB values were calculated using national datasets, and may fail to incorporate sub-national differences in deforestation. Using Kenya as an example, a study of regional analyses highlights significant differences in estimated fNRB, ranging between 23% to 70%⁴ (Figure 4). While some projects in this sector group are at national scale, many are smaller scale and implemented in rural, remote, and marginal areas, where fNRB estimates may vary considerably.
2. Use of Demonstrably Renewable BiomassTo calculate the fNRB, two key inputs are required: the shares of renewable and non-renewable woody biomass in total biomass consumption. In the original CDM/UNFCCC defaults, it was assumed that the only renewable biomass nationally was ‘Demonstrably Renewable Biomass’ (DRB). What qualified as DRB was based on a strict set of criteria, for example, biomass originating from protected areas such as wildlife reserves and national parks. This assumption meant that all biomass in other land areas that were not explicitly sustainably managed was non-renewable. This approach overestimated fNRB because it underestimated the renewable biomass available, and did not properly consider fuelwood sourced from land other than forests. Fuelwood can be sourced from a variety of carbon pools and land covers, such as deadwood, home gardens, live fences, and residues from timber and/or agriculture.⁵ Therefore, changes in household fuel use may not have a material impact on national deforestation.
There have been attempts to produce more accurate measures of fNRB, most prominently the Woodfuel Integrated Supply/Demand Overview Mapping (WISDOM) model and Modelling Fuelwood Sustainability Scenarios (MoFuSS). WISDOM models woodfuel supply and demand dynamics at a more regional level,⁶ using remote-sensing data and geographic information systems (GIS). It takes into account socioeconomic, legal, and topographic variables that may affect access to woody biomass. MoFuSS is similar, yet incorporates a dynamic temporal model which simulates supply and demand patterns over time. These approaches are likely to provide a more accurate estimate of fNRB, which aligns with BeZero’s framework for assessing over-crediting risk.
However, estimating fNRB on a project-specific level can be time consuming, challenging, and expensive, so understandably projects may opt to use default values instead. Since 2020, the country-specific defaults originally approved by the CDM and UNFCCC have no longer been permitted for use in CDM projects.⁷ The new default adopted by the CDM is 30%, regardless of the host country, which was determined using global results from the WISDOM model.⁸ This represents a 57% difference compared to the average country-specific default (87%), with the lowest default being 65%⁹ and the highest being 100%.¹⁰ This indicates some acknowledgement by the Standards Body that previous default values were too high, and supports our analysis of over-crediting risk for projects using such defaults.
Methods and tools to determine fNRB are designated by a project’s chosen methodology under the Standards Body under which credits have been issued. Assessing the quality of a carbon credit project therefore begins with understanding the methodology it uses, and how it is applied. We explore how this applies to the Household Devices sector group below.
Projects under the CDM can opt to use this 30% default, or calculate fNRB values for their country or project area using the CDM’s TOOL30. Projects accredited under the Gold Standard are also required under the most recent methodologies to use the default 30% or TOOL30, as Gold Standard’s default values for Peru, Bolivia, Colombia, Honduras, and Guatemala expired in 2021.¹⁴ TOOL30 uses a similar equation to the original fNRB formula, however it assumes that renewable biomass can originate outside protected areas, and incorporates other sustainably managed types of land cover and geographically remote areas where deforestation is unlikely to occur from household consumption.
We note that for projects rated by BeZero, fNRB values remain high even with the use of TOOL30. Of those that have used TOOL30, the largest observed difference in TOOL30 calculated versus country-default values has been only 0.6%, and such projects are maintaining overestimated fNRB values. For example, across five Kenyan projects rated by BeZero, three rely on the CDM/UNFCCC default value for Kenya, one uses its own calculation, and one used TOOL30. Yet, all reported fNRB values range between 91.5 - 97%. This indicates that further analysis of the sources of data and how they are applied within TOOL30 is needed for such projects.
This is where BeZero’s bottom-up analysis goes further to interrogate project claims and assess over-crediting risks at a regional level. Our team examines the quality of data sources used, and the inputs and assumptions a project uses in calculating their fNRB to assess whether the claimed value is appropriate and conservative, all of which can skew the final results of fNRB calculations. As higher fNRB values lead to more claimed emission reductions, examination of these parameters are critical.
Interestingly, this 30% default value is yet to be adopted by any of the projects that have been assigned a BeZero Carbon Rating. Across Household Devices projects, BeZero has maintained the view that conservative global fNRB values sit closer to the range of 20-30%. In terms of the actual impact on emission reductions, one study of 286 carbon credit projects found that the difference between WISDOM values and what projects report has led to the overestimation of emission reductions by around 41-59%.¹⁵
Default country-level fNRB values have previously relied on unrealistic assumptions about sustainable woodfuel harvesting. In our opinion, using such default values can create significant over-crediting risk, and the use of more conservative values can temper this risk.
Although new calculation methodologies are being introduced, the underlying assumptions and data sources of such fNRB values require further interrogation to ultimately determine a reliable estimate of Household Devices projects’ climate impact.
⁴Values taken from Bailis et al., 2015
⁹Lowest CDM/UNFCCC fNRB default applies to Jamaica
¹⁰Highest CDM/UNFCCC fNRB default applies to Bahrain & Comoros
¹³The only approved fNRB under standardised baseline approach is for Myanmar