The carbon costs of global wood harvests

The paper addresses the significant yet often overlooked carbon emissions associated with global wood harvesting. While tree growth absorbs carbon, various carbon accounting methods lead to widely varying estimates of the net carbon impact of wood harvesting. Many approaches suggest low, zero, or even negative emissions because they offset carbon losses from harvests with carbon sequestration from the growth of broader forest areas. This offsetting is inappropriate as this growth would occur independently of new harvests, driven by factors like agricultural abandonment, forest recovery, and climate change. Some methods also overlook the time value of carbon, only considering gross annual emissions without accounting for the eventual regrowth of harvested forests. The study's central research question is to accurately quantify the carbon costs of global wood harvests, factoring in the time value of carbon using a time-discounting approach, and assess the potential for reducing these costs as a climate change mitigation strategy. The study's importance lies in providing a more accurate assessment of the climate impact of wood harvesting, informing policies aimed at reducing greenhouse gas emissions, and promoting sustainable forest management practices.

The paper reviews existing methods for accounting for the greenhouse gas (GHG) effects of forest harvests, highlighting inconsistencies and biases across lifecycle analyses of wood products, national GHG reporting, and scientific analyses of land-use change emissions. Lifecycle analyses often deem wood use 'carbon neutral' if forests are managed sustainably, essentially ignoring the initial carbon release from harvest. National reporting uses netting approaches, allowing countries to claim credit for regrowth after agricultural abandonment or previous harvests, potentially masking the impact of new harvests. Scientific papers estimating land-use change emissions sometimes also report net effects, obscuring the separate impact of new harvests. The reviewed literature reveals a strong regional bias in these accounting methods, potentially portraying temperate countries' harvests as benign while highlighting the costs of harvests in tropical countries. Critically, the paper cites numerous scientific sources demonstrating the illogical nature of offsetting new harvest emissions with forest regrowth that would occur regardless of the harvests. Although some studies report gross emissions, they are deemed inadequate as they don't consider regrowth. The existing literature is characterized by varying and often inaccurate methodologies for estimating the carbon cost of wood harvesting, creating the need for a more robust and accurate approach.

The study employs a novel global forest carbon model, the carbon harvest model (CHARM), to estimate the carbon costs of wood harvests between 2010 and 2050 under various scenarios. CHARM tracks carbon shifts among different storage pools (live vegetation, roots, slash, wood products, and landfills) and calculates atmospheric carbon change over time. The model's key innovation is incorporating time discounting, valuing carbon losses and gains differently depending on the year they occur. This is achieved by applying a 4% discount rate to emissions and removals over time, resulting in 'harvest-year equivalent emissions'— effectively translating the future flow of emissions and removals into a present-day value. The choice of a 4% discount rate aligns with a middle-ground estimate of a constant social cost of carbon and a reasonable rate of return on capital, and is consistent with existing policies valuing land-use change emissions. The model projects future wood consumption by country for four product categories (long-lived, short-lived, very-short-lived wood fuel, and very-short-lived industrial products) using a fixed-effects model based on historical relationships between consumption, population, GDP, and time. The model then estimates annualized carbon costs under seven scenarios varying wood supply and demand, considering factors such as secondary forest regrowth, plantation expansion, productivity improvements, harvest efficiency, and wood fuel demand reductions. The model also assesses potential substitution benefits—reductions in production emissions when wood replaces concrete, steel, or propane gas—but importantly notes that these benefits do not offset the biogenic emissions from wood harvest itself. The study further uses 'clear-cut equivalents' to estimate harvested land area.

The study finds that forest harvests between 2010 and 2050 will likely have annualized carbon costs of 3.5–4.2 Gt CO₂e yr⁻¹, comparable to common estimates of annual emissions from agricultural expansion. These costs are largely driven by existing wood demand (78%), with rising demand accounting for the remainder. The analysis shows that industrial wood and wood fuel each account for about half of the total carbon costs. Scenarios with lower carbon costs include those with reduced wood fuel demand or increased plantation growth rates. Substitution benefits, from reduced production emissions when wood replaces other materials, are estimated at 0.8–0.9 Gt CO₂e yr⁻¹, but these do not negate the biogenic emissions from wood harvest. The study estimates that 756–855 million hectares of land would be harvested across different scenarios, with strategies such as increased plantation growth and reduced wood fuel demand leading to significant reductions in harvested area. Sensitivity analyses show that the main finding of around 3–5 Gt CO₂e yr⁻¹ in decadal effects is robust across various model parameters. The study also notes that its estimates are likely conservative due to the omission of soil carbon losses and indirect effects of forestry. Finally, the findings show that the results are relatively insensitive to changes in discount rates or time horizons, only exhibiting significant variation when using zero discounting over 100 years.

The study's findings directly address the central research question by providing a robust estimate of the carbon costs of global wood harvests. The significant carbon costs of 3.5–4.2 Gt CO₂e yr⁻¹, comparable to agricultural expansion, challenge the common assumptions of carbon neutrality or even carbon benefits associated with wood harvesting. This highlights a large and often overlooked contribution to climate change. The relatively low sensitivity of the results to discount rate variations indicates that the short-to-medium term climate impacts are substantial. The study's contribution to the field is its development of a more accurate and comprehensive accounting method for the carbon costs of wood harvests, addressing critical shortcomings of existing approaches. The identification of the potential to reduce these costs through mitigation strategies, such as reducing wood fuel demand or increasing plantation growth rates, offers valuable insights for policy-makers and forest managers. The discussion emphasizes the importance of considering the absolute carbon emissions from wood harvest, regardless of alternative human activities that might occur in a counterfactual scenario. The study implies that reducing wood harvests could be a significant climate change mitigation wedge.

This study demonstrates that global wood harvests incur substantial and often underestimated carbon costs, approximately 3.5–4.2 Gt CO₂e yr⁻¹, comparable to emissions from agricultural expansion. The model, incorporating time discounting, offers a more precise accounting of carbon flows than previous methods. The relatively insensitivity to discount rates underscores the importance of near-term emission reductions. Reducing wood harvest represents a significant underappreciated climate change mitigation opportunity. Future research could focus on refining estimates of soil carbon losses and indirect effects of forestry, and on exploring the economic and social implications of various forest management and wood use reduction strategies.

The study acknowledges several limitations. The estimates of soil carbon loss and recovery are uncertain, potentially underestimating the overall carbon costs. Indirect effects of forestry, such as those from road building, are not included. The model's projections of future wood demand rely on assumptions about population growth, economic development, and technological change, which could affect the accuracy of the results. The clear-cut equivalent used for land area estimation could lead to some imprecision.