Over-reliance on land for carbon dioxide removal in net-zero climate pledges

The paper examines how national climate pledges (NDCs and LT-LEDS) translate expected carbon dioxide removal (CDR) into land demands, addressing the gap in current assessments that quantify CDR only in tonnes of CO₂ without specifying the forms and land implications of these removals. Because land-based approaches currently dominate CDR and land is under competing pressures from biodiversity conservation, food production, and livelihoods, the study asks what aggregate land area and land-use changes are implied by existing national pledges and how these demands are distributed over time and geography. The authors situate the work within the Paris Agreement’s goal to balance sources and sinks in the second half of the century, noting that land-use change is a leading driver of biodiversity loss and that large-scale land-based CDR (e.g., afforestation, BECCS) can both aid mitigation and exacerbate ecological and social risks. They highlight concerns that reliance on future CDR may delay near-term emissions reductions and emphasize the importance of transparency about land requirements embedded in pledges. The study aims to provide an initial, comparable estimate of land demands implied by national pledges up to end-2023 and to inform governance discussions on which CDR methods are intended, by whom, where, and at what scale and timing.

The authors summarize prior assessments showing that current NDCs and mid-century pledges are inadequate to meet Paris goals, while many countries plan substantial CDR including land sinks. Literature highlights that land-use change drives biodiversity loss and can undermine Indigenous and local livelihoods, with large-scale afforestation and BECCS carrying significant ecological and social risks. Modelling studies project large areas for reforestation and bioenergy crops in 1.5 °C pathways, raising feasibility and justice concerns due to unprecedented land transformation rates. There is also discussion in the literature of how discursive reliance on CDR can induce a ‘spiral of delay,’ weakening near-term mitigation. Restoration pledges under other conventions (UNCCD, CBD, Bonn Challenge) reach around 1 billion ha and only partially overlap with climate mitigation pledges, underscoring coordination and transparency challenges. The IPCC has flagged key governance questions around which CDR methods to deploy, at what volumes, by whom, where, and when.

Sample and sources: The study reviewed LT-LEDS and/or NDCs submitted to the UNFCCC up to end-2023 for 194 countries; the EU and its 27 Member States were assessed as one bloc. Three countries had not submitted NDCs. For Brazil, the 2015 NDC was assessed per government statements to reinstate it. Where UN submissions omitted CDR but other official statements existed, those were used. The longest-term pledge available (LT-LEDS as proxy for net-zero/2050 targets) was prioritized, assuming near-term land-based CDR is encompassed by longer-term pledges. Pledge typology and conversion to land area: Pledges were categorized into (1) direct land-area pledges (hectares, acres, km²), (2) indirect area pledges (e.g., % forest cover increase or number of trees), and (3) emissions-based pledges (removals in MtCO₂ or % of emissions). Direct areas were taken as stated. Indirect pledges were converted to area using FAO land cover/use datasets and, for tree-count pledges, ecoregion-specific tree densities from Crowther et al. Emissions-based pledges were converted to land area using default IPCC removal factors by activity type and climate domain. CDR activity classification: Land management activities mentioned in pledges were classified into seven IPCC-aligned activity types based on sequestration potential: primary forest (protection; excluded as non-anthropogenic CDR), old secondary forest (restoration of degraded forests), young secondary forest (reforestation/expansion, mixed species), plantations (commercial/monocultures), mangroves (restoration/expansion), agroforestry (trees in croplands/regenerative agriculture), silvopasture (trees/rangeland restoration), and bioenergy crops (for BECCS). Activities were grouped as restoration (no land-use change) versus reforestation/plantations/energy crops (land-use change). DACS was included as novel CDR with no land area assumed. Assumptions and parameters: Forestry removal factors were primarily from Harris et al.; agricultural removal factors from the IPCC; BECCS land area was estimated assuming dedicated energy crops using country-specific median yields (Li et al.) and a 60% conversion efficiency (Vaughan et al.). No separate land area was estimated for bioenergy demand without CCS. Protection/avoided deforestation and primary forest removals were not counted as additional CDR. Only future land-based CDR was included (areas already counted toward targets were excluded). Temporal and geographical analysis: The authors aggregated implied land areas by target year to estimate temporal scaling (e.g., to 2030, 2050, 2060) and summarized contributions by country to assess geographical concentration. Uncertainty and sensitivity: Uncertainty was propagated for emissions-based and indirect pledges from the uncertainties in removal factors (forestry, agriculture, bioenergy) and tree densities. Direct area pledges were assumed to have zero uncertainty. A sensitivity test using a global average removal factor increased total land area by 8.4% (117 Mha); including below-ground biomass increment decreased total by 8.9 Mha (<1%). The uncertainty analysis is conservative, as it excludes uncertainties from pledge interpretation and biome/activity classification. Data availability: Datasets are deposited on figshare (doi:10.6084/m9.figshare.24080472), with supplementary details provided in the article’s Supplementary Information and Source Data.

- Aggregate land demand: 990 million ha (range 892–1087 Mha) of additional land would be required to meet pledged CDR from 2020–2060. This exceeds the land area of the United States (983 Mha) and is roughly two-thirds of global cropland (1561 Mha in 2020). - Land-use change vs restoration: 435 Mha (395–475) would involve land-use change (reforestation, plantations, energy crops); 555 Mha (466–644) would be restoration of degraded forests, agricultural lands, or wetlands (no land-use change). - CDR typology breakdown (Table 2 indicative values): Old secondary forest (restoration): ~483.7 Mha; young secondary forest (reforestation): ~345.6 Mha; plantations: ~28.1 Mha; bioenergy crops for BECCS: ~61.0 Mha; silvopasture: ~48.5 Mha; agroforestry: ~22.5 Mha; mangroves: ~0.41 Mha. DACS assumed no land area. - Temporal scaling: About 211 Mha additional land is relied upon by 2030, scaling to ~990 Mha by 2060. The implied rate of land-use change for reforestation is up to 13 Mha per year from 2020–2050 (assuming linear scale-up with ~50 Mha after 2050). - Comparison to scenarios and history: The 13 Mha/yr implied rate is unprecedented historically and comparable to average rates in 1.5 °C model pathways by mid-century (which often assume ~17 Mha/yr). Scenario medians suggest ~322 Mha forest expansion and ~199 Mha energy cropland by 2050 in 1.5 °C pathways. - Geographical concentration: Four countries account for over 70% of the implied land area: Russia (35.4%), Saudi Arabia (20.2%), United States (10.9%), Canada (5%); the top 10 countries account for ~85%. Saudi Arabia’s pledge includes ~40 billion trees (~200 Mha) largely outside its borders. The US strategy implies ~54 Mha reforestation plus ~54 Mha for BECCS at the upper end of its modelled CDR. India’s pledge implies ~24 Mha of additional forest; China’s 25% forest cover by 2030 implies ~19 Mha expansion from 2020; the EU includes significant restoration under its LULUCF regulation; Australia’s estimate is inflated by reliance on BECCS and international credits. - Disproportionate shares for low-income countries: Many low-income and vulnerable nations (particularly in Africa) pledge large shares of national land area to CDR; two countries (Dominica, Equatorial Guinea) exceed 100% when combining direct area and emissions-based pledges. - BECCS specifics: Quantified BECCS land (five countries) totals ~61 Mha; pledges lack detail on feedstocks. Land needs are sensitive to yields and conversion efficiency; alternative assumptions or use of wastes could change estimates, but pledges as stated imply large land areas. - Sensitivity/uncertainty: Using a global average removal factor increases land needs by 8.4% (117 Mha); adding below-ground biomass reduces totals by ~8.9 Mha (<1%).

The analysis demonstrates that national climate pledges embed very large expectations for land-based CDR, addressing the research question by quantifying the forms and scale of land implied rather than only tonnes of CO₂. The findings indicate that the pledged reliance on land—especially via reforestation and prospective BECCS—is high enough to risk delaying near-term emissions cuts, substituting impermanent or reversible land sinks for largely permanent fossil CO₂ emissions. The scale and speed of implied land-use change approach or exceed historically unprecedented rates and mirror contested assumptions in 1.5 °C modelled pathways, raising feasibility, governance, and justice concerns. The geographical concentration among major fossil producers and the high relative burdens in low-income countries suggest potential inequities and risks of land rush dynamics, with possible impacts on food security, biodiversity, and land-dependent livelihoods. The heavy emphasis on reforestation and plantations heightens risks to biodiversity and can drive conversion of non-forest ecosystems (e.g., grasslands), while restoration activities also carry risks of access restrictions and dispossession depending on governance contexts. BECCS pledges are highly uncertain regarding feedstocks and supply chains; large-scale deployment could compete with agriculture unless mitigated by yield gains, governance, or dietary shifts. The study underscores the need for transparency in pledges—clear specification of land areas, CDR methods, locations, and timelines—and for separating emission reduction from removal targets and land-sector from energy-sector goals. Prioritizing restoration of degraded lands over new plantations or forest expansions can reduce land-use conflicts and biodiversity harms. Overall, the results argue against escalating land-based CDR ambitions as a substitute for stronger near-term emissions reductions.

National pledges imply roughly 1 billion hectares of additional land for CDR by 2060, with over 40% requiring land-use change for reforestation, plantations, or energy crops, and the remainder focused on restoration. This level and pace of land transformation are unprecedented and risk undermining food security, biodiversity, and social equity, particularly in the Global South, while not guaranteeing delivery of pledged removals within stated timelines. The authors call for: (1) greater ambition in near-term emission reductions to reduce reliance on CDR; (2) explicit disclosure of land areas and CDR methods in pledges; (3) separation of emissions reduction and removal targets, and of land and energy sector components; and (4) prioritization of degraded land restoration over tree planting and forest expansion. Future work should improve pledge transparency, coordination across environmental conventions, and spatially explicit analyses to assess trade-offs and governance needs for equitable, realistic land-based mitigation.

- Coverage gaps: No quantifiable CDR estimates for 51 countries, including some large forested nations; many countries lack 2050 LT-LEDS, so land reliance could be larger as more long-term pledges are submitted. - Pledge vagueness and assumptions: Net-zero targets are often non-specific, requiring assumptions to map emissions-based pledges to land areas, potentially over- or underestimating land needs. - BECCS uncertainty: Pledges lack detail on biomass feedstocks, yields, and conversion efficiencies; land requirements are highly sensitive to these parameters and to potential use of waste feedstocks. - Scope exclusions: Protection/avoided deforestation and primary forest removals were excluded; no assessment of land area for bioenergy without CCS; only future additional land-based CDR counted. - Biophysical simplifications: Removal factors are largely aboveground biomass; below-ground increments change totals marginally; climate-driven changes in yields and sequestration (e.g., CO₂ fertilization effects, productivity impacts) are uncertain and not fully captured. - Uncertainty analysis: Propagated only from denominator uncertainties (removal factors, tree densities); does not include uncertainties from pledge interpretation, biome/activity classification, or spatial allocation, likely underestimating total uncertainty. - Delivery timelines: Biomass growth lags mean pledged removals may not materialize within stated timeframes, affecting comparability with target years.