Introduction

Carbon accounting aims at setting the base for calculations on the offset of carbon emissions in difference harvest and management approaches. Two conceptual approaches are proposed in the literature to calculate carbon emissions (GOFC-GOLD 2008). The stock-based approach estimates the difference in carbon stocks in a particular pool at two points in time and is often applied in setting national-wide baseline degradation rates based on national forest inventories (GOFC-GOLD 2008). In contrast, the gain-loss approach estimates the net balance of additions to and removals from a carbon pool for different carbon pools (GOFC-GOLD 2008). We opted for implementing a gain-loss approach for two reasons. First, cross-sectional data sets on diameter increments and diameter distributions were measured as part of the participatory forest inventory. These data are required for the gain-loss approach but are not sufficient for the stock-based method. Second, it is possible to compare sustainable forest management schemes and business-as-usual degradation in terms of income for local people and carbon stocks but altering the rate of forest loss.

The model is inspired by CO2FIX (Schelhaas et al. 2004, Masera et al. 2003) in terms of structure, modules and simplicity but deviates in how it treats tree mortality, timber harvest and financial feasibility calculations. The projection of carbon stocks is not as comprehensive as a process-based vegetation model (e.g. LPJmL, Sitch et al. 2003). This is because the model uses simplified mechanisms of forest regeneration, growth, mortality and harvest in line with the limited data available in tropical and subtropical countries (Schelhaas et al. 2004, Masera et al. 2003). The model must be prepared to simulate the impact of demand for forest products (fuelwood, timber in different diameter-classes etc.) on the development of the total carbon stock. A major prerequisite in REDD studies is to precisely define forest degradation and deforestation. We use the definition of forest degradation based on IPCC (2003), extended by Griscom et al. (2009: 7) as the “direct, human-induced reduction in the forest carbon stocks from the natural carbon carrying capacity of natural forest ecosystems which persists for a specified performance period and does not qualify as deforestation”. Therefore, the objectives are to estimate the natural carbon carrying capacity and the historical forest degradation of natural forest ecosystems based on calculating the carbon stock and flows in different carbon pools by means of a simulation model with feedback mechanisms of forest growth and forest degradation. The carbon stock in different carbon pools are projected according to different policy scenarios: business-as-usual degradation and sustainable forest management.

Simulation model and scenario setup

The simulation model consists of six modules – regeneration, growth, mortality, harvest, carbon budgeting and dynamic investment calculation. We loop the model over N number of years, with forest regeneration, growth and natural mortality taking place in each year.

• Forest regeneration is a function of the change in number of trees per hectare due to mortality and harvest events
• Forest growth is based on mean annual diameter increment from observed current annual increment data over different diameter classes from the forest inventory
• Mortality due to senescence is approximated by a fixed maximum diameter which can be translated to the ratio of current biomass to potential maximum biomass per tree
• Tree mortality prior to mortality due to senescence is implemented as function of forest density in three diameter cohorts
• The business-as-usual scenario harvests trees above a threshold growing stock per hectare without a predefined harvest interval
• Sustainable forest management-based scenarios give priority to restoring the natural carbon carrying capacity with harvesting taking place at predefined harvest intervals
• The carbon budgeting module compiles the carbon emissions due to harvesting and carbon stored in each of five pools: aboveground biomass, belowground biomass, soil, litter (aboveground, belowground and foliage dead organic matter)
• Dynamic investment calculations estimates the net income under different harvesting conditions for a predefined REDD project lengths. This is done with carbon costed at different dollar prices

A spin-up phase of 200 years is needed to bring forest regeneration, growth and natural mortality into equilibrium, where the change in volume increment per hectare is close to zero. Without sample plots in undisturbed natural forest, the potential maximum biomass stock has been estimated from sample plots in village 6 in 2004. This is an estimate of the forest capacity without human intervention. Placing a ‘harvest shock’ on undisturbed growth causes the forest to try and restore the equilibrium condition. We shock the model in equilibrium with the observed forest degradation between 2004 and 2009. Thus, all harvest scenarios start from the degraded forest status in 2009 while forest use continues according to business-as-usual or under sustainable forest management scenarios.

Baseline scenario

The business-as-usual practice of wood harvest follows governmental regulations on the quantity of harvestable biomass. There is a prescribed minimum threshold of biomass volume of 130 m3 per hectare to be maintained on the stock. In addition, the critical minimum threshold of timber harvest is 50 m3 to warrant cost-efficient harvest and transport to saw mills. Thus the volume of aboveground biomass is used as criterion to define sustainably-managed forests, but leaves diameter distribution out of considerations. There are two major arguments against this simple harvesting protocol. First, harvesting diameter classes from 30 cm to 40 cm of major marketable tree species disturbs the natural diameter distribution, changing the species composition and disrupting income from timber sales. Second, the business-as-usual scenario alters the canopy layers and post-harvest mortality leading to changes in the ecosystem such as surface runoff and soil stabilization.

Sustainable forest management (SFM) scenario

Sustainable forest management takes harvesting and ecosystem services into account by selective logging over a range of diameter classes. This limits the impact of changing the diameter distribution and canopy layers. To evaluate the feasibility of these scenarios the revenue lost in reduced timber sales must be compared to the revenues gained from carbon credits for carbon sequestered.

In contrast to the baseline scenario there is no prescription on the minimum threshold of harvestable timber. Harvest is prescribed to be the arithmetic mean of gross annual volume increment in fixed periodic harvest intervals. Only trees with a diameter of 25 cm and above qualify for being harvested. Increment and Forest Growth

Biomass growth is determined by tree diameter increment in each year of simulation. The mean annual increment (MAI) has been calculated as arithmetic mean from trunk discs of randomly selected cut trees gathered during the forest inventory in Village 6. The last five years were ignored since harvesting will have affected stand density and in turn MAI. According to these measurements an MAI of 0.5 cm*y^-1 for the time period before 2004 is estimated. This value is assumed to apply to the entire forest area.

Carbon budgeting

The aboveground biomass of each tree in kg dry matter was estimated with diameter-based biomass expansion factors in line with FAO methodology. Observed values for aboveground mass in kg CO2 per diameter class were used to fit biomass expansion functions which were the same for all of the villages. The harvest residues, i.e. tree crown wood, foliage and stumps, remain in the area and enter the litter pool. The total carbon pool is the sum of all components of the aboveground biomass, the belowground biomass, the soil carbon and the litter carbon due to harvest and senescence.

The dynamic soil carbon model Yasso (Liski et al. 2005) is used to get a rough estimation for the carbon pool in the soil. The model describes decomposition and dynamics of soil carbon in well-drained soils. The current version is calibrated to describe the total stock of soil carbon without distinction between soil layers. The model can be applied for both coniferous and deciduous forests. It has been tested to describe in a wide range of ecosystems from arctic tundra to tropical rainforest. There is no interaction between the Yasso soil model with the biomass growth model.

Financial feasibility calculation

The financial analysis compares costs and revenues based on the total carbon accumulated in different scenarios over time. The question pertains to the given price of CO2/tonne to make a REDD project financially feasible. The REDD project is treated as any other capital investment. The opportunity costs of foregone timber extraction are considered the minimum payable to the resource owner to agree to reduce harvest activity. There is no reason however why the payment for sequestering carbon should not exceed this minimum. Since there is reason to expect uncertainty in the carbon price we analyze with carbon prices of 10USD, 25USD and 50USD throughout the duration of the project.

Different cost types, i.e. setup costs (forest management planning) treated as fixed costs, operating costs (monitoring costs, cutting and hauling costs) and natural resource tax on harvested timber volume are defined from local data and expert estimates. However, several other setup costs types, i.e. feasibility study, preparing communities for participation, setting up payment scheme and operating costs, i.e. administration costs have not been included. Outputs

The output is thematically divided into biological and economic outputs. Biological outputs comprise, inter alia, carbon pools for each 5-year time step for the number of time steps simulated. Financial feasibility results are produced in separate files. Output is summarized in separate files for every management scenarios in csv-format for the ease of further processing.

You can download the final report on the Carbon Model from here