Net-zero emission targets for major emitting countries consistent with the Paris Agreement

The study examines when major emitting countries are projected to reach domestic net-zero CO₂ and total GHG emissions consistent with 1.5 °C and 2 °C pathways. Using national-level outputs from integrated assessment models (IAMs), the authors assess how these phase-out years compare to the global average, how definitional choices (e.g., land-use accounting, allocation of negative emissions) affect timing, and which country characteristics explain cross-country differences. The work aims to inform policymakers in setting credible, nationally appropriate net-zero targets.

The scenario results align with broad findings from the IPCC Special Report on 1.5 °C regarding the timing of global net-zero CO₂ and GHG emissions. Prior studies are referenced showing near-zero energy CO₂ emissions before 2050 for Brazil and the USA based on national models (e.g., Schaeffer et al.). Equity-based analyses (e.g., Robiou du Pont et al.) provide alternative, fairness-oriented timelines that can imply earlier phase-out years for countries with lower per-capita emissions or developing economies, complementing the cost-optimal approach used here.

- Scenario set: The analysis uses nationally disaggregated outputs from six IAMs: AIM, IMAGE, MESSAGE-GLOBIOM, POLES (also referred to as POLIS in the text), REMIND-MAgPIE, and WITCH. Scenarios target 1.5 °C and 2 °C (≥66% probability), under globally cost-optimal mitigation (equal marginal GHG price across countries). - Phase-out year estimation: For each country and scenario, the year when CO₂ (fossil CO₂ and total CO₂) and total GHG emissions (Kyoto gases, including land use) reach net zero is identified. Extrapolated emissions data are used where needed to avoid bias in calculating phase-out years and their differences across sensitivity cases. - Definitions and sensitivity analyses: Four influential definitional choices are explored: 1) Harmonization of LULUCF: Model projections are harmonized to countries’ reported LULUCF inventory data by adding the 2010 inventory–model absolute difference to projected land-use emissions; impacts on phase-out years are quantified. 2) Allocation of negative emissions (BECCS): Default model allocation credits negative emissions to the carbon-storing country. A sensitivity reallocates BECCS negative emissions to the biomass-producing country; effects on national phase-out years are quantified ex-post using biomass sourcing shares. 3) Global warming potentials (GWPs): Effects of using different 100-year GWP values (e.g., IPCC AR5 vs AR4) on GHG phase-out timing are tested. 4) Equity vs cost-optimal: National phase-out years from cost-optimal scenarios are compared with ranges derived from equity-based rules from the literature. - Explanatory factors: Fifteen candidate variables (e.g., CCS uptake share in 2050, afforestation/deforestation potential, transport CO₂ share, non-CO₂ share, GDP per capita, land cover shares) are screened for redundancy (R² criterion, pairwise correlations, visual inspection). Multiple linear regression over combinations of variables selects a parsimonious model explaining variance in phase-out years across ten countries. - Regression dataset: Projections from POLES and IMAGE are used for regression since they provide full coverage across all countries under both temperature targets, ensuring balanced records per country. - Additional policy context: Supplementary analysis links long-term neutrality timelines to near-term milestones (e.g., 2030 reductions vs 2015; 2050 reductions), referencing NDC alignment. - Data and code: Scenario data are available via the CD-LINKS explorer; policy indicators via the Global Stocktake tool. The R scripts used for figures are available on GitHub and Zenodo.

- Global timing: In 1.5 °C scenarios, global GHG emissions reach net zero between 2050–2070; in 2 °C scenarios, after 2080. CO₂ reaches net zero earlier: 2045–2060 for 1.5 °C and 2065–2080 for 2 °C. - Country timing relative to global average: • Earlier than global average: Brazil and the USA (for both total GHGs and CO₂ in many scenarios); Japan earlier for GHGs in cost-optimal cases. • Later than global average: India and Indonesia (notably later); Canada and Turkey often later for certain gases/targets; China and the EU near the global average. • For fossil CO₂ only (excluding land use): Brazil, Indonesia, Japan and the USA tend to reach net zero earlier than the global average in 2 °C scenarios (Canada and the USA in 1.5 °C scenarios). Countries where land use is a net source (e.g., Indonesia) see later phase-out for total CO₂ than for fossil CO₂ only, while land-use sinks (e.g., Canada) can advance total CO₂ phase-out relative to fossil CO₂. - Near-term milestones (2 °C scenarios, relative to 2015): Many countries peak by 2030; 2030 reductions range from 12% (India) to 36% (Japan, Canada, Indonesia). By 2050, reductions range from 52% (Brazil) to 72% (USA); under 1.5 °C, USA reductions can reach up to 90% by 2050. - Sensitivity to definitions: • LULUCF harmonization: Using inventory-based LULUCF generally advances net-zero GHG timing in all countries except Brazil (where inventory shows higher land-use emissions than model LUC), with particularly large impacts for countries with significant or uncertain land-use fluxes (e.g., Indonesia). • BECCS allocation: Reallocating negative emissions to biomass producers brings earlier phase-out years for Brazil, Indonesia, Canada, India and Russia; it delays phase-out for biomass importers such as the EU, Japan and Turkey. • GWPs: Updating 100-year GWP values has negligible effects on phase-out years. • Equity framing: Equity-based allocations would imply earlier phase-out years for lower per-capita emitters and developing economies (e.g., Brazil, Indonesia) compared to cost-optimal allocations. - Drivers of cross-country variance: Multiple linear regression indicates that negative-emissions-related potentials (CCS uptake share and afforestation capacity) are dominant predictors of earlier phase-out years. Additional significant factors include the shares of transport CO₂ (associated with earlier phase-out), non-CO₂ emissions (later), and GDP per capita (later in cost-optimal allocations). Forest share is less robust in parsimonious models. - Sectoral contributions at phase-out: Remaining emissions often include CH₄ and N₂O; in some models F-gases remain significant in China and Japan. Transport contributes to residual CO₂ in nearly all countries (except Russia). Negative CO₂ is predominantly from BECCS in most countries; Brazil relies more on afforestation. Some models project negative industry CO₂ in Brazil, Russia, Canada and to a lesser extent the EU. - Country-specific notes: Brazil combines relatively high non-CO₂ shares with substantial negative emissions potential, enabling early phase-out. Countries with late phase-out (e.g., India, Indonesia) show larger residual or baseline growth and lower negative emissions potential.

The analysis addresses when major emitters can achieve domestic net-zero emissions under globally cost-optimal 1.5 °C and 2 °C pathways and why their timelines differ. Results show that timing is highly sensitive to land-use accounting choices and the allocation of negative emissions, underscoring the need for clear, consistent definitions in target-setting and international reporting. The dominant role of negative-emissions potential (CCS and afforestation) explains why countries like Brazil and the USA tend to achieve earlier neutrality, while higher shares of hard-to-abate non-CO₂ emissions and higher GDP per capita (under cost-optimal allocation) correlate with later phase-out. Policy relevance is high: several announced net-zero targets (China 2060 carbon neutrality; EU 2050 net-zero GHG; Japan 2050 net-zero GHG; USA 2050 net-zero GHG) align broadly with modelled domestic cost-optimal pathways for 1.5–2 °C, though aggregate NDCs remain insufficient. The findings inform national target design, including specification of coverage (CO₂ vs all GHGs), land-use accounting, and roles for negative emissions, and provide context for Article 6 mechanisms and ITMOs to reconcile cost-effectiveness with equity.

This paper advances understanding of national net-zero timelines consistent with Paris Agreement temperature goals by providing country-level neutrality years from IAM scenarios and identifying key drivers of cross-country differences. It shows that Brazil and the USA can reach net zero earlier than the global average, while India and Indonesia are later, with China and the EU near the global average, and that definitional choices (LULUCF accounting, BECCS allocation) materially affect timing. Negative-emissions capacity (CCS and afforestation) is the primary driver of earlier phase-out years, alongside sectoral composition and non-CO₂ shares. Future research should expand scenario coverage to more countries and model types (national, sectoral, macroeconomic), test alternative metrics (beyond GWP) and consumption-based accounting, conduct deeper single-model diagnostics, and compare with countries’ long-term strategies to improve political realism. Incorporating social-science insights on feasibility and policy implementation, and clarifying international rules (e.g., Article 6) will further support credible, equitable national net-zero target-setting.

- Scenario assumptions reflect globally cost-optimal mitigation and may not capture equity-based national choices or political feasibility; the analysis excludes endogenous equity considerations. - The scenario set (developed circa 2016–2018) assumes idealized global policy rollout and may diverge from recent political realities in some countries (e.g., Brazil), and current NDCs are insufficient. - Large model spread exists for some countries (notably Brazil and India), reducing certainty in specific phase-out years. - Land-use emissions differ substantially between national inventories (managed land balance) and model LUC estimates (human-induced vegetation change), leading to sensitivity in results when harmonized. - Country coverage varies by model; regression analysis relies on two models (POLES and IMAGE) for full country coverage, which may limit generalizability of statistical inferences. - Code for IAMs is not publicly available; while data and figure scripts are accessible, full model reproducibility is constrained.