The clean development mechanism and community forests in Sub-Saharan Africa: reconsidering Kyoto’s “moral position” on biocarbon sinks in the carbon market

Abstract

Negotiations of the Kyoto Protocol reached what has been called a moral position on biocarbon sinks which saw important limitations on their use in the Clean Development Mechanism (CDM), the Protocol’s main carbon offset system. After outlining this moral position, this article examines the consequences of these limitations on the viability of community forest participation in the CDM through a case study of three community forests in West Africa. Results suggest that there is significant carbon mitigation potential from forest conservation, reforestation as well as from improved fuelwood cookstoves at the community level. Yet under the current rules of the CDM, little of this overall carbon mitigation potential is able to be realized. Using qualitative research methodologies, it was learned that community respondents showed a pragmatic, yet cautious interest in the CDM while also emphasizing a need for land-use flexibility. The paper closes with a political discussion of the “‘moral position” on biocarbon sinks in the carbon market and concludes with policy recommendations for biocarbon sinks, in both the CDM and REDD, in the post-Kyoto climate change regime.

Appendix 1: Quantitative carbon mitigation potential methodology

Quantitative carbon mitigation potential was assessed for reforestation, forest conservation (i.e. avoided deforestation), improved cookstoves and electricity generation. Field visits in all three communities permitted information to be collected through participant observation, forest walks and the consultation of forest management plans. Land-use zoning and forest management planning had been undertaken by the Ekuri and Tinto communities but not yet in Tali at the time of fieldwork, which was in the process of establishing a community forest (Akumsi et al. 2005).

1.1 Reforestation methodology

Reforestation potential was based upon data provided by E. Asaah (pers. comm.) at ICRAF, Cameroon, on growth rates over the first 2–4 years of two key agro-forest species (Njansang—Rhicininodendron heudelotii; Plum—Dacroides edulis) and their combination under variable planting densities (100, 204 and 400 trees/ha). In order to assess reforestation potential within the short term, the combined growth data of Njansang and Plum at a low density of 100 trees/ha were converted to CO₂ and subject to linear regression to estimate CO₂ sequestration up to their fifth year (i.e., 2012, the end of the Kyoto Protocol’s first commitment period). The low-planting density was selected as it is most compatible with continued agro-forest activity. These data were then entered into a spread sheet designed by the World Bank’s Biocarbon Fund to estimate CO₂ sequestration (Noble et al. 2005) and applied over the period 2008–2012 over areas of farm-fallowlands in Ekuri.

1.2 Avoided deforestation methodology

The CO₂ equivalent content of standing forest was determined using 2003 IPCC Good Practice Guidelines for LULUCF (Anonymous 2003). This drew on above-ground biomass (dry matter) presented in Table 3A.1.4: Cameroon = 131 tonnes/ha; Nigeria = 184 tonnes/ha. Below-ground biomass was derived from the root–shoot ratio for primary tropical forest in Table 3A.1.8: 0.22–0.33. Emission factors for the burning of biomass are complex (Fearnside 2000a). In this case, emissions factors were derived from parameters for tropical forests in the 2006 IPCC Guidelines for National Greenhouse Gas Inventories (Aalde et al. 2006: Table 2.5): 1,580 ± 90 g/kg DMB for CO₂; 6.8 ± 2.0/kg DMB for CH₄; 0.2/kg DMB for N₂O (DMB = Dry Matter Burnt). CH₄ and N₂O were further transformed by multiplying by their global warming potential, 21 and 310, to convert into CO₂ equivalent. A conservative estimate (using lower bounds of all parameters), this led to an emission factor for the burning of tropical forest at 216 tCO₂e/ha for emissions from the combustion of above-ground biomass and 51.3 tCO₂e/ha for below-ground (267.7 tCO₂e/ha total) if all the forest were burned for the Cameroonian estimate. Since above-ground biomass estimates in Cameroon were larger than those in Nigeria (Anonymous. 2003: Table 3A.1.4), the Cameroonian data were applied to all standing forests to be conservative. It should be noted that the above calculations did not include an assessment of carbon content in understorey vegetation, deadwood and soils, which might increase total biomass to 450 tonnes/ha (Kotto-Same et al. 1997: 248–249) report 204 tC/ha based on a carbon content of 0.45). In other words, the carbon stock estimates here might underestimate CO₂e losses by 30% (N. Bird, pers. comm.).

However, not all forest stock is burned during swidden agriculture, and the rate of regrowth is important. To estimate this, the chronosequence (including 25-year-old cocoa plantation) described by Kotto-Same et al. (1997: 249) was used. Note that discrepancies in their biomass estimations, described above, precluded the use of their actual figures. However, the relative change in carbon content was used to estimate change in CO₂ equivalent content over the chronosequence. Note that below-ground biomass was not observed to change significantly. Results are presented in Table 3 below. Total system CO₂e/ha emission released, at the appropriate stage of the chronosequence, was used when estimating emissions avoided by forest conservation in Ekuri and Tinto.

Putting the conversion factor discussed earlier with empirical data from Kotto-Same et al. (1997), These data demonstrate that conversion from standing tropical forest (which initially retained 267.7 tCO₂e/ha) by way of swidden agriculture results in the initial release of 254.9 tCO₂e/ha, nearly all of it due to a reduction in above-ground biomass. It should be noted that over time, as the area is laid fallow, vegetative regrowth accumulates CO₂ so that 17 years after swidden agriculture clearance a net emission of only 77.6 tCO₂e/ha has resulted. Carbon stocks would be replenished again to original 267.7 tCO₂e levels after approximately of 25 years. Kotto-Same also compared the carbon content of the standing forest against a 25-year-old cacao farm, which retain only about 179.4 tCO₂e/ha. Cocoa can grow in the shade and a few large trees are generally productive for 25 years. Therefore, converting standing tropical forest into a cocoa farm results in the release for up to 25 years of 88.3 tCO₂/ha (Table 3).

Table 3 CO₂-equivalent stock of standing forest derived from Kotto-Same, Woomer, Appolinaire and Louis (1997: 249)

1.3 Improved fuelwood cookstove methodology

Fuelwood consumption was only empirically measured at Tali as part of improved fuelwood cookstove assessment project (Akumsi et al. 2005; Purdon 2005). Because of similar cooking methods were observed across all three communities, fuelwood cookstove results were extrapolated post hoc to populations of Ekuri and Tinto. A total of 41 households were visited (n = 41) in Tali-Bara in 2005 in order to conduct semi-structured interviews, primarily with women, to discuss and quantify fuelwood consumption (for a detailed methodology consult Purdon (2005)). It should be noted that the survey was made in the dry season (March and April) when fuelwood is plentiful. Domestic cooking was the activity most frequently reported during interviews in Tali, while one-third of households interviewed reported they cooked occasionally for market. Most women collected fuelwood from one-half to two-hours per day.

Average daily consumption of fuelwood was measured at approximately 8–9 kg fuelwood/day per household, mostly collected from household farms. Many of these farms were recently burned (1–2 years), though they had initially been established as long as 15–25 years prior. Upscaling daily consumption across the community (approximately 124 households) and on an annual basis, Tali was estimated to consume 362–409 tonnes fuelwood/year. The lower-boundary, conservative estimate of fuelwood consumption (362 tonnes fuelwood/year) was used in estimating all remaining calculations.

Fuelwood data gathered in Tali-Bara were converted into CO₂e via two methods. The first was estimated directly from fuelwood (Meth-1) while the second has been approved by the CDM Executive Board (CDM-EB 2008), referred to hear as Meth-CDM. The first, Meth-1, was based on an assessment of the emissions associated with the combustion of fuelwood directly, using the same 2006 IPCC emission factors for assessing avoided deforestation above (Aalde et al. 2006: Table 2.5): 1,580 ± 90 g/kg DMB for CO₂; 6.8 ± 2.0/kg DMB for CH₄; 0.2/kg DMB for N₂O (DMB = Dry Matter Burnt). The lower boundary of the emission factors was used in order to be conservative. Lastly, it was assumed that there was near complete combustion of the fuelwood. One tonne of fuelwood was estimated to result in the emission of 1.69 tonnes CO₂e.

The second, Meth-CDM, used the small-scale methodology AMS I.E: Switch from Non-Renewable Biomass for Thermal Applications by the User which requires fuelwood to be converted into a fossil fuel baseline (CDM-EB 2008). This is accomplished by converting fuelwood into its energy equivalent (the methodology provides an IPCC default for wood fuel, 0.015 TJ/tonne) and then further converting this into emissions via an emission factor of 71.5 tCO₂/TJ for Kerosene, 63.0 tCO₂/TJ for Liquefied Petroleum Gas). To achieve maximum emissions most comparable to fuelwood, the emission factor for Kerosene was used.

A locally designed cookstove developed by Dr. George Eyabi (pers. comm.) at IRAD-Batoke was found to reduce fuelwood consumption by 38.5% versus a traditional three-stone hearth. If improved cookstoves were adopted by all households in the Tali community, this would result in a reduction of 139 tonnes fuelwood/year.

Finally, note that in order to extrapolate Tali-Bara results to Ekuri and Tinto, the same fuelwood consumption and associated emission reductions from the improved cookstove via Meth-1 and Meth-CDM were extrapolated. Tali-Bara results were based on fuelwood consumption at the household level. While the exact number of households was not determined in Ekuri nor Tinto, the number of households could be estimate from the population/household ratio derived empirically for Tali (4.69 persons per household). This was applied back to the population estimates of Ekuri and Tinto to arrive at a household number of 1,279 and 426, respectively.

1.4 Electricity generation/solar power methodology

A number of electric generators were encountered in each community, running on diesel fuel. In Ekuri, there was little evidence of the use of electricity, though reportedly there were a number (less than 30) of individual 5 kWe generators used intermittently. Tinto had one 25–30 kWe diesel generator that had been donated to the community (which ran 4 h a day, 5 days a week) and a weekly fuel consumption estimate of 430 l was provided by its caretaker. Note that the generator was not in operation during field visits. In addition, both Tinto and Tali had a number of smaller generators present. While a formal survey was not undertaken, a generous estimate would suggest that each community in Cameroon possessed an additional 15 individual 5 kWe generators. To offer the best possible contrast to carbon mitigation potential under biocarbon sinks, it was assumed that these generators were also each running for four hours a day, five days a week and would be replaced by a carbon neutral technology such as solar power (see Table 4 for details). Such a scenario would generate the greatest amount of carbon credits possible for comparison with the carbon mitigation potential of biocarbon sinks.

The carbon mitigation potential for electricity generation was estimated by assuming replacement of electricity generation via solar power, reducing all current emissions from electricity generation. Emissions from electricity generation were based on the small-scale methodology AMS 1.A. Electricity Generation for User and the energy baseline is the fuel consumption of the technology in use or that would have been used in the absence of the project activity to generate the equivalent quantity of energy For the large 25–30 kWe diesel generator at Tinto, emissions could be derived from weekly fuel consumption estimates provided by its caretaker (430 l at the time of fieldwork, 340 l when new). The generator was reported to be used 5 days a week for 4 h a day (1,040 h/year). Emissions were derived by first assuming the density of diesel to be 0.87 kg/l (IOR Energy 2003 reports a range of 0.85–0.88 kg/l). Diesel emission factors for CO₂, CH₄ and N₂O were then applied from Table 1.4 of the 2006 IPCC Guidelines for National Greenhouse Gas Inventories (Garg et al. 2006) and then multiplied by each GHG’s global warming potential. Obtaining fuel consumption rates for the smaller ~5 kWe generators was not possible in the field and needed to be estimated post hoc. The general rule of thumb for fuel consumption is 7% of the rated generator output (GeneratorJoe 2008). For 5 kWe, this is 1.32 l/h. See Table 4.

Table 4 Emissions from electricity generation