International shipping in a world below 2°C

The study addresses how international shipping can align with global climate goals of well below 2°C and preferably 1.5°C. Shipping is a hard-to-abate sector with rising transport demand and limited remaining efficiency gains; without intervention, emissions could grow to 1.1–2.1 GtCO2/yr by 2050 at current carbon intensity. The 2018 IMO strategy targets at least a 50% reduction in GHG emissions by 2050 vs. 2008, but the sector’s role in deep decarbonisation pathways remains uncertain due to technological inertia (long ship lifetimes, prevalent compression ignition engines), need for new bunkering infrastructure, and strong linkages to the global energy system. The paper situates shipping within economy-wide mitigation strategies using IAMs to evaluate feasible emissions trajectories and fuel transitions under different global carbon constraints, and compares these with the IMO2050 target.

Prior research on shipping decarbonisation has largely relied on sectoral models, with IAMs historically offering limited detail on maritime transport. Recent IAM developments have begun incorporating shipping-specific demand and mitigation options, enabling integrated analysis of system-wide interactions, including energy supply, land use, and carbon removal. The literature highlights limited efficiency potentials, the importance of low-carbon fuels (biofuels, synthetic fuels, hydrogen/ammonia), infrastructure and technology readiness challenges, and competition for biomass between transport fuels and negative emissions (BECCS). Existing scenarios reviewed by IPCC WGIII rely on CDR to compensate hard-to-abate emissions; however, the degree to which shipping can decarbonise within broader system constraints and diverse fuel portfolios has been underexplored in a multi-IAM setting.

The authors conduct the first multi-IAM comparison focused on international shipping using six global models with varying shipping-sector granularity: COFFEE, IMAGE, PROMETHEUS, and TIAM-UCL (higher technological resolution for maritime fuels and propulsion) and IMACLIM-R and WITCH (lower resolution, focusing on a subset of alternatives). All scenarios adopt SSP2 socioeconomic assumptions. Three scenario groups are assessed for 2020–2100: (1) NDC (continuation of 2015 pledges without an explicit global emissions cap), (2) C1000 (global carbon budget of 1,000 GtCO2, likely below 2°C), and (3) C600 (600 GtCO2, slightly above 1.5°C), with peak-budget dynamics disallowing net negative emissions after reaching net zero. International shipping emissions are unconstrained sectorally, allowing endogenous allocation of mitigation across sectors. Models differ in shipping demand representation, technology options, and fuel production pathways. Energy carriers are harmonized into categories for comparability: conventional petroleum fuels; oilseed-based fuels; drop-in synthetic biofuels; other drop-in synthetic fuels (power-based); alcohols and gases (fossil/bio/synthetic); hydrogen and ammonia; and direct electricity. Additional regional analysis for the EU (PRIMES-Maritime) is provided in the Supplementary Information. Model descriptions detail structures: process-based optimization (COFFEE, TIAM-UCL), CGE (IMACLIM-R), simulation (PROMETHEUS), intermediate-complexity process model (IMAGE), and hybrid macro-energy optimization (WITCH). Efficiency improvements in shipping are generally exogenous; fuel switching and demand responses to carbon pricing are endogenous where applicable.

• Under NDC trajectories, international shipping emissions stabilize or increase through the century, driven by rising transport activity and limited efficiency gains. Projections for 2050 fall broadly in the order of 0.6–1.4 GtCO2/yr, increasing further by 2070–2090 in some models.
• With global carbon budgets, emissions decline despite rising activity via efficiency improvements and especially fuel switching:
  • C1000 (1,000 GtCO2): 2050 international shipping emissions span ~0.3–1.2 GtCO2/yr. Relative to 2008: COFFEE −26%, IMAGE −34%, PROMETHEUS −65% (exceeding IMO2050 and IEA SDS), IMACLIM-R −9%, TIAM-UCL +49% (delayed fuel transition with post-2050 catch-up to −58% by 2070), WITCH +16% (limited fuel options lead to mitigation shifted to other sectors).
  • C600 (600 GtCO2): deeper reductions by 2050. COFFEE −43%, IMAGE −59% (≈IMO2050/SDS), PROMETHEUS −88% (≈IEA NZE). IMACLIM-R similar to C1000 (biofuel potential saturated), TIAM-UCL −2% (faster transition), WITCH +2%.
• Long-term stabilization: In models running to 2100, emissions tend to stabilize in latter half-century under C1000/C600 as BECCS deployment rises after 2040; however, large-scale BECCS feasibility is uncertain (land and water constraints), implying higher decarbonisation pressure on shipping if limited.
• Energy demand: Total maritime fuel demand converges to ~10–16 EJ/yr by 2030; diverges by 2050 (7–23 EJ/yr) and 2070 (12–28 EJ/yr) due to activity and efficiency assumptions. Models with demand response and stronger efficiency (IMAGE, PROMETHEUS) show 7–8 EJ/yr in 2050 (25–50% lower than their NDC cases), while high-activity/low-efficiency cases (e.g., IMACLIM-R, TIAM-UCL NDC) reach 20–23 EJ/yr.
• Fossil share and model granularity: Shipping retains a higher fossil share than other sectors even in deep mitigation. Model detail strongly influences fossil reliance:
  • C1000 2050: high-resolution shipping models show 46–67% fossil in shipping (vs 34–65% in global primary energy); low-resolution models show 64–96% (vs 46–55%).
  • C600 2050: high-resolution 18–63% (vs 34–52%); low-resolution 58–88% (vs 28–47%).
• Fuel portfolios: A diverse set of low-carbon fuels is adopted. Pathways with very high renewable power deployment (e.g., TIAM-UCL) favor green ammonia (≈9 EJ/yr in 2070; ≈18 EJ/yr in 2090). Biomass-rich pathways (IMACLIM-R, IMAGE) favor bio-based synthetic liquids (bio-alcohols, FT fuels), but biomass competes with BECCS for negative emissions, limiting fuel availability. COFFEE and PROMETHEUS show diverse mixes with early adoption of vegetable oils and renewable alcohols in the 2040s, and later dominance of green ammonia (~3 EJ/yr) and lignocellulosic biofuels (~2 EJ/yr) in COFFEE; PROMETHEUS sees lignocellulosic bioenergy exceed 2 EJ/yr by 2050 in C600.
• LNG plays a minor role in carbon-constrained scenarios due to limited mitigation potential compared to truly low-carbon fuels; some NDC cases show 2–5 EJ/yr LNG by 2050. TIAM-UCL sometimes uses more natural gas under budgets to bridge rapidly growing demand when ammonia supply lags.
• Overall, models with broader representations of low-carbon options achieve deeper maritime decarbonisation by 2050, occasionally surpassing IMO2050 targets, while limited-option models keep higher fossil shares and shift mitigation burdens to other sectors/BECCS.

The multi-IAM analysis demonstrates that meeting Paris-consistent carbon budgets requires international shipping emissions to plateau or decline despite rising activity, primarily via accelerated deployment of low-carbon fuels. The findings address the central question by showing that decarbonisation compatible with below-2°C is feasible when shipping is integrated into economy-wide mitigation, with technology choice influenced by broader energy system evolution (renewables scale-up, biomass availability, BECCS deployment). Crucially, the richness of available maritime fuel options within a model materially affects outcomes: greater technological diversity enables deeper cuts and reduces dependence on CDR, whereas limited fuel portfolios maintain high fossil shares and delay decarbonisation. The sector’s intrinsic inertia (long-lived assets, engine constraints) underscores the need for early investments in alternative fuels, propulsion systems, and bunkering infrastructure to align with mid-century goals. The role of LNG is limited under stringent budgets, while drop-in biofuels and renewable alcohols support near-term transitions and ammonia and synthetic fuels become essential post-2030/2040.

This study provides the first coordinated, multi-IAM assessment of international shipping under Paris-aligned carbon budgets, showing that substantial emission reductions—up to 88% by 2050 in some pathways—are achievable primarily through fuel switching to a diverse set of low-carbon options. Models with detailed maritime technology/fuel representation indicate greater decarbonisation potential than models with limited options, highlighting the importance of technology diversity to limit residual emissions and reliance on CDR. Near-term deployment of drop-in biofuels and renewable alcohols can accelerate the transition, with ammonia and synthetic fuels becoming pivotal beyond 2030–2040. Given sectoral inertia, early investments in low-carbon fuels, engines, and infrastructure are critical. Future research should refine technological detail in IAMs for shipping, assess supply-chain constraints and sustainability of biomass and e-fuels, evaluate infrastructure roll-out and safety standards (e.g., ammonia), and explore policies that efficiently coordinate global maritime decarbonisation with broader energy system transitions.

• Large-scale deployment of BECCS underpins mid/late-century stabilization in many scenarios, but is uncertain due to land and water constraints; tighter BECCS limits would increase decarbonisation pressure on shipping.
• Some models have limited shipping-sector granularity, potentially underestimating decarbonisation potential and over-relying on mitigation in other sectors or CDR.
• Efficiency improvements are often exogenously specified; real-world diffusion may differ.
• Biomass availability and sustainability constraints, and electrofuel production scale-up, pose upstream uncertainties not fully resolved here.
• The analysis harmonizes SSP2 and specific carbon budgets; results may vary under different socioeconomic pathways or tighter budgets where some models fail to solve.