There has been extensive debate about whether the sustainable use of forests (forest management aimed at producing a sustainable yield of timber or other products) results in superior climate outcomes to conservation (maintenance or enhancement of conservation values without commercial harvesting)1, 2, 3, 4, 5, 6, 7, 8. Most of the relevant research has relied on consequential life-cycle assessment (LCA), with the results tending to show that sustainable use has lower net greenhouse-gas (GHG) emissions than conservation in the long term1, 2, 3, 4, 5. However, the literature cautions that results are sensitive to forest- and market-related contextual factors: the carbon density of the forests, silvicultural and wood processing practices, and the extent to which wood products and forest bioenergy displace carbon-intensive alternatives. Depending on these issues, conservation can be better for the climate than sustainable use1, 6, 7, 8. Policy institutions are another key contextual factor but, so far, they have largely been ignored1, 2, 3, 4, 5, 6. Using a case study on the Southern Forestry Region (SFR) of New South Wales (NSW), Australia, we show how policy institutions can affect the assessed outcomes from alternative forest management strategies. Our results highlight the need for greater attention to be paid to policy institutions in forest carbon research.
Institutions are generally defined as norms that structure human behaviour and social interactions9, 10. In the current context, the phrase ‘policy institutions’ is used to refer to the rules and procedures adopted by governments and inter-governmental organizations concerning GHG mitigation and accounting.
Borrowing from LCA nomenclature, we classify relevant policy institutions as macro, consequential or attributional11, 12, 13. Macro policy institutions are those concerning policy objectives and principles. Consequential policy institutions are those that affect substantive outcomes; they provide incentives (inducements, penalties or information) for policy actors to behave in ways that affect emissions and removals. Attributional policy institutions are those that assign responsibility for emissions and removals between jurisdictions and other relevant actors (Table 1).
Table 1: Types of policy institutions relevant to mitigation strategies for forests.
For each of the eight scenario sets, we modelled emissions and removals from the six relevant sinks and sources: on-site forest carbon; harvested wood products; landfill; fossil fuel emissions from forest management, transport, and wood processing; emissions from product substitution; and net avoided emissions through bioenergy production. Supplementary Tables 1 and 2 summarize the coverage of the sinks and sources.
On-site forest carbon.
The Southern Forestry Region (SFR) consists of a mix of forest types dominated by Eucalyptus and Corymbia spp., including C. maculata, E. muelleriana, E. pilularis, E. sieberi, E. obliqua, E. fastigata, E. cypellocarpa and E. delegatensis14, 26, 27, 28. Most of the estate is regrowth and mature forest, with only a small fraction (0.5%) being high conservation value old growth (Supplementary Fig. 1)29, 30.
Mirroring the Australian Government’s approach to modelling public native forests31, the SFR forest estate was modelled using the Tier 2 capabilities of FullCAM (version 3.30.1; ref. 32). For modelling purposes, the region was divided into its three management sub-regions (South Coast, Tumut and Eden) and representative FullCAM forest plots were devised for each sub-region. Details of the gross area and net harvestable area of the sub-regions are provided in Supplementary Table 3 (refs 26, 27, 28).
The representative FullCAM plots were based on the ‘medium dense eucalypt forest’ and ‘tall dense eucalypt forest’ plots used in the Australian Government’s public native forest model31. The medium dense eucalypt forest plot provided the basis for the South Coast and Eden plots, the tall dense eucalypt forest plot provided the basis for the Tumut plots. Adjustments were made to these base plots to account for the assumed silviculture practices, basic density and above-ground biomass yields in the SFR.
Silviculture practice assumptions.
Eighteen representative plots were developed that broadly reflect current silvicultural practices and forest types in the SFR (Supplementary Table 4)14, 26, 27, 28, 29, 30. For the South Coast, we assumed 64% of the estate was harvested by way of modified single tree selection (STS), with the remainder subject to Australian group selection (AGS; ref. 26). Based on state forestry agency data, 12% of the estate was assumed to be thinned at 30 years26. For Tumut, we divided the net harvestable area into two broad forest types—alpine ash and mountain hardwood—and assumed harvesting in both was by way of STS or AGS. No thinning was assumed to occur in the Tumut sub-region27. To reflect the impact of environmental restrictions, in each rotation, we assumed 10% of the net harvestable area in the South Coast and Tumut sub-regions was not harvested33. For Eden, the estate was broken into regrowth and older multi-aged forest, with the latter comprising 15% of the sub-region28. The majority of existing multi-age forest was assumed to be harvested over the first two decades of the projection period, thereby converting it into managed regrowth. Two-thirds of all regrowth forest was assumed to be thinned at 30 years, with a subsequent regeneration harvest at 70 years. The remainder was assumed to be subject to a single regeneration harvest at 60 years.
All plots in each sub-region were assumed to have the same slash proportions (Supplementary Table 5)31, 34. However, the proportion of above-ground biomass assigned to slash was assumed to be lower in the bioenergy scenarios to reflect the use of sub-pulp grade logs for bioenergy. In the bioenergy scenarios, all branches, bark, leaves and roots were assumed to be left on-site to maintain soil fertility14, 35, 36.
Basic density.
The basic densities in the base plots were adjusted to reflect the forest types in the sub-regions (Supplementary Table 6)14, 31, 34, 37. All other parameters, including yield allocations to tree components, carbon and turnover percentages, and debris breakdown percentages, were assumed to be the same as those in the relevant base plot.
Above-ground biomass yields.
The above-ground biomass yields in each plot were modelled using the equation:38, 39
ABY is the above-ground biomass yield (bone dry metric tonnes), Age is the stand age in years, α is the maximum attainable above-ground biomass (upper asymptote of the curve) and β and γ determine the shape of the curve (growth rate to the asymptote). The standard parameters for α, β and γ used in the plots are provided in Supplementary Table 7. Adjustments were made to these parameters to account for the impact of thinning (increased growth rates of residual trees and impeded new growth).
The parameters for the plots in Supplementary Table 7 were developed iteratively to ensure the modelled sub-region roundwood removals matched sustainable yield forecasts published by the Forestry Corporation of New South Wales (NSW; refs 26, 27, 28). The published sustainable yield forecasts do not cover all log categories. To address this, we used historical data to calculate a ratio between the published log categories and total roundwood removals17, 18, and then applied the ratio to the forecast yields. The resulting correlation between the modelled roundwood removals and adjusted sustainable yield forecasts is shown in Supplementary Fig. 2.
Harvested wood products and landfill.
Harvested wood products and landfill carbon stocks and emissions were modelled using an integrated version of the Australian Government’s models31, 40. The inputs for the integrated product/landfill model were derived from the FullCAM outputs.
Over the period 2003–2011, 71% of roundwood removals from the SFR were pulplogs, 26% were sawlogs and 3% were other logs (for example, poles, girders, landscaping and sleeper logs; Supplementary Fig. 3)17, 18. In the no bioenergy scenarios, these proportions were used as the basis for assigning roundwood removal outputs to log categories. In the bioenergy scenarios, we assumed sub-pulp grade logs (which would otherwise have been left as slash) and 50% of pulplogs were used for bioenergy14, 35, 36. The proportion of total roundwood removals (including removals for bioenergy) assigned to each log type is provided in Supplementary Table 8.
The processing destination fractions for logs and wood waste were derived from the Australian Government’s harvested wood products model, with adjustments made to account for regional industry characteristics (Supplementary Data File)31, 40. The destination fractions provided the basis from which end-products were assigned to the product pools contained in the harvested wood products model (Supplementary Table 9). The maximum age and decay rates for the product pools are summarized in Supplementary Table 1031, 40. Exported wood chips were not modelled within the Australian Government’s harvested wood product model paper pool (pool 1). Consistent with the international accounting rules, they were modelled using the IPCC first-order decay function, with a paper default half-life of two years19. The destination fractions for losses from the product pools are provided in the Supplementary Data File.
The Australian Government’s landfill model is based on the IPCC Tier 2 first-order decay (FOD) model21, 31. The key parameters of the model are the fraction of degradable organic carbon in each individual waste type (DOC); the rate of decay assumed for each individual waste type (decay function ‘k’); the fraction of degradable organic carbon that dissimilates through the life of the waste type (DOCf); the methane correction factor (MCF); the methane recovery rate (proportion of methane captured for flaring and energy generation); and the oxidation factor (the proportion of methane that oxidizes before reaching the surface of the landfill)21, 31. The inputs for the landfill model were derived from the HWP model in tonnes of carbon (tC), making the DOC value redundant. Details of the remaining parameters are provided in Supplementary Table 11.
Fossil fuel emissions from transport, processing and management.
The fossil fuel emissions associated with harvesting, hauling, processing and transporting wood and waste products to relevant markets, and the fossil fuel emissions associated with forest establishment and management, were calculated using data from industry and government sources41, 42, 43. Details of the energy and emission factors, and transport distance assumptions are provided in the Supplementary Data File.
Product substitution.
Since the mid-1990s, hardwood log production from NSW public native forests has fallen by 50% (Supplementary Fig. 4)44, 45, 46, 47, 48. Similar trends have been seen across Australia, with falls in native hardwood log production being experienced in all relevant states20, 22, 24, 49. The decline in production, both in NSW and Australia-wide, is attributable to a combination of market-related factors—particularly weak demand growth and increased competition in domestic solid wood product marke
Andrew Macintosh. Rethinking forest carbon assessments to account for policy institutions[J]. Nature Climate Change,2015-06-29,Volume:5:Pages:946;949 (2015).
