Integrated assessment modeling of a zero-emissions global transportation sector

The study addresses how the global transportation sector—responsible for over 20% of CO₂ emissions—can achieve deep decarbonization compatible with limiting warming to 1.5 °C. Decarbonizing transport is challenging due to oil dependence, dispersed emissions, slow response to carbon pricing, and the technical difficulty of decarbonizing long-distance modes like aviation and maritime shipping. Prior work often treats modes in isolation, focuses on road transport, or lacks integrated feedbacks with other sectors, leaving gaps in understanding sector-wide strategies and interdependencies. This paper uses an integrated assessment framework to evaluate technology pathways and timelines for phasing out fossil fuels across all transport modes, especially aviation and shipping, and examines system-wide implications for fuels, land use, energy, and emissions under 1.5 °C-consistent constraints. The purpose is to identify viable technology portfolios, quantify emissions outcomes, and understand cross-sector tradeoffs under varying levels of transport decarbonization ambition.

The paper situates its contribution within several strands of literature: (1) sector-specific and industry scenarios for aviation and shipping that lack full economy-wide integration; (2) integrated modeling studies that focus regionally, aggregate transport, prioritize road vehicles, or include limited mitigation options for aviation and shipping; (3) analyses of deep decarbonization in aviation and freight without detailing relative contributions of low-carbon fuels. Recent net-zero commitments by aviation and shipping stakeholders underscore the need for integrated modeling of aggressive decarbonization across modes. The study builds on and extends prior integrated assessments by explicitly representing advanced options for aviation and shipping (e.g., FT biofuels and e-fuels) and by evaluating technology-specific contributions and cross-sectoral impacts under stringent climate constraints.

Model: Global Change Analysis Model (GCAM) version 6.0 linking energy, water, land, and climate with detailed transport representation. Runs executed on PNNL’s HPC cluster. The Hector climate module provides temperature projections. Scenario design: Four scenarios with SSP1 population and GDP (for primary cases): - Reference: Continuation of current trends, no climate policy, base GCAM transport technology assumptions. - 1.5 °C low transport tech: Moderate adoption of electric and hydrogen technologies; no new fossil-fuel vehicle sales by 2060 (passenger cars/buses/trucks) and 2090 (aviation/shipping); no sector-specific fossil phase-out beyond these sales bans. - 1.5 °C medium transport tech: Fast integration of electric and hydrogen; no new fossil-fuel vehicle sales by 2050 across modes; aviation/shipping fossil liquids replaced by bio-/e-fuels by 2100. - 1.5 °C high transport tech: Most ambitious adoption; no new fossil-fuel sales by 2030 (aviation/shipping sales ban for fossil liquids applies via blending mandates); fossil liquids in road/rail phased out by 2050; aviation/shipping fossil liquids fully replaced by bio-/e-fuels by 2050. Demand-side assumptions (held constant across the 1.5 °C scenarios): Reduced transport service demand via lower income elasticities, increased ridesharing (higher LDV load factors to +25% by 2050), and higher public transit preference (SSP1 value-of-time). Income elasticity adjustments are summarized in Table 2 (e.g., long-haul aviation from 1.0 to 0.75; other passengers to 0.8; international shipping 0.4 to 0.3; other freight 0.75 to 0.5). Climate constraint: A carbon emissions constraint applied to fossil fuel and industry CO₂ from 2025 ensures end-of-century warming below 1.5 °C, consistent with IPCC AR6 Category 1 pathways (no/low overshoot). Land-use CO₂ is priced (fraction of the fossil carbon price) but not constrained. Non-CO₂ GHGs are unconstrained but included in forcing. Technology and fuels: In addition to existing GCAM options (electricity, hydrogen, biofuels), the authors add synthetic hydrocarbon e-fuels produced from DAC-derived CO₂ and hydrogen (four DAC-to-fuels configurations varying electricity and H₂ sourcing: grid plus delivered H₂; grid plus on-site electrolysis; dedicated wind; dedicated solar powering both DAC and electrolysis). Aviation/shipping biofuels are constrained to FT fuels from cellulosic biomass for sustainability and suitability. A global bioenergy use cap is imposed: 100 EJ by 2100 (primary scenarios) to mitigate land competition and direct biomass to high-value uses. For aviation (hydrogen combustion turbines) and shipping (hydrogen fuel cells), liquid fuel blending mandates enforce minimum shares of FT biofuels/e-fuels over time; for road/rail, liquid fuels are fully phased out by 2050 (high) or 2100 (medium). Upstream generation: Electricity and hydrogen production technologies are endogenously determined, with electrolysis options linked to grid/renewable supply; CCS/BECCS options compete under carbon pricing and bioenergy constraints. Cost and comparison metrics: Break-even carbon prices for low-carbon shipping and aviation technologies are computed as cost differences per unit service divided by direct emissions differences, using reference scenario cost assumptions (energy + non-energy) and assuming zero direct emissions for electric/hydrogen. DAC cost ranges come from GCAM assumptions (Supplementary Table 1). Comparisons are made with AR6 database scenarios for transport subsectors. Sensitivity analysis: Variants of the high-ambition transport scenario include: alternative SSPs (SSP2–SSP5 socioeconomics); restoring default income elasticities; reducing price elasticities by 20%; and a stricter bioenergy cap (70 EJ by 2100). For SSP3/SSP5, negative emissions budget ceilings are increased by 50% to ensure feasibility under the 1.5 °C constraint. Data/code availability: GCAM v6.0 and inputs/outputs are available via GitHub/Zenedo; analysis scripts are archived on Zenodo.

- All 1.5 °C scenarios achieve economy-wide deep mitigation to keep warming below 1.5 °C by 2100; transport’s contribution varies by ambition. - Reference vs 1.5 °C high: Transport direct emissions eliminated by 2050 in the high scenario; total (direct+indirect) transport emissions reduced by 99% relative to 2020. Reference scenario transport emissions rise >40% from 2020 to 2100. - Cumulative emissions: From 2050–2100, residual transport emissions are 39 Gt CO₂ (medium) and 157 Gt CO₂ (low). Offsetting with DAC would cost about $7–14 trillion (medium) and $27–54 trillion (low) (2020$; cost assumptions in Supplementary Table 1). Across 2020–2100, high vs reference reduces direct transport emissions by 675 Gt CO₂; high vs low reduces by 245 Gt CO₂ (~60% of the IPCC AR6 two-thirds 1.5 °C carbon budget). Non-CO₂ and air pollutants (e.g., CO, NOx) decline markedly. - By mode (high vs low, cumulative): Largest absolute savings from freight trucks (78 Gt CO₂) and passenger cars/trucks (70 Gt CO₂). Largest percent reductions: long-haul aviation (77%) and short-haul aviation (75%). Aviation+shipping together cut 83 Gt CO₂ and account for 34% of transport-sector savings. - Relative to AR6 database: High and medium scenarios are below median transport emissions in 2050 and 2100; AR6 scenarios with detailed mode splits generally do not reach zero emissions in aviation/shipping until 2075 or later, making this study’s high case more ambitious. - Technology contributions (high scenario): • Electrification dominates most modes by 2050+, except international shipping and long-haul aviation. • International shipping relies mainly on hydrogen by 2050; alternative liquid fuels are important earlier. Domestic shipping is >80% electric by 2050. • Long-haul aviation depends on alternative liquid fuels (biofuels/e-fuels), supplying >88% of service from 2050; e-fuels remain <16% of the alternative liquid mix due to higher costs. • Short-haul aviation is more balanced: ~56% electric service by 2050; hydrogen and alternative liquids split the remainder. • Road and rail are highly electrified; buses use more hydrogen (≈44% hydrogen, 56% electric post-2050). - Cost-competitiveness: Break-even carbon prices for electric/hydrogen shipping and short-haul aviation are ~135–496 2020$/tCO₂ by 2050, comparable to projected DAC costs (172–351 2020$/tCO₂). Long-haul electric/hydrogen aviation requires >1300 2020$/tCO₂, indicating significant additional support/innovation needs. - Service and fuel use: By 2030, non-fossil fuels provide 19% of transport fuel in the high scenario (>2× the low scenario). Carbon pricing suppresses aviation/shipping service in lower-ambition cases; higher ambition mitigates service reductions (2–7% higher cumulative service for aviation/international shipping in high vs low). - Cross-sector impacts: Lower transport ambition forces deeper mitigation elsewhere; in the low scenario, industry reaches zero emissions by 2100. High vs low cumulative emissions differences: industry 128 Gt CO₂, electricity 74 Gt CO₂. The high scenario reduces CCS needs (less BECCS for power/H₂; less industrial CCS early-century) and lowers total sequestration by 0.2–0.7 Gt CO₂ per period from 2030. - Fuel demands and upstream generation: In 2050, transport consumes 17+ EJ H₂ (69% of all H₂) in high, 9.4 EJ (47%) in medium, 5.5 EJ (28%) in low. From 2025 onwards, all FT biofuels and e-fuels are used by aviation/shipping (aviation majority). By 2050, 99% of electricity and ≥98% of hydrogen are low-carbon. In high, added H₂ demand is mainly met by electrolysis using grid electricity; H₂ from BECCS drops to 5% (1.3 EJ) vs 17% (3.4 EJ) in low. - Sensitivity results: Cumulative transport emissions differences vs high-ambition baseline are <8 Gt CO₂ (<7%). However, fuel requirements vary: combined H₂+electric transport use in 2100 is >50% higher (income elasticity adjustment) and >15% higher (price elasticity adjustment) than the high baseline. SSP3/SSP4 reduce total transport energy (higher carbon prices boost H₂ competitiveness, notably in international shipping). SSP2 shifts away from biofuels toward H₂/electricity. SSP5 increases use of all fuels, with e-fuels in 2100 >3× the high baseline. Under a stricter bioenergy cap (70 EJ), transport biofuel use in 2100 is ~25% lower (~3 EJ), offset mostly by more electricity and e-fuels.

The findings show that ambitious, early transport decarbonization materially advances 1.5 °C goals and eases mitigation burdens on other sectors, especially industry. High ambition reduces reliance on negative emissions (DAC/BECCS) and lowers cross-economy sequestration. Within transport, rapid electrification is the cornerstone for road and rail, while hydrogen and advanced liquid fuels are critical for hard-to-electrify modes. Long-haul aviation’s decarbonization leans heavily on biofuels under the study’s sustainability and cost assumptions; however, technology advances or policy support could elevate hydrogen’s role and/or increase e-fuel uptake under high economic growth or constrained bioenergy. Ambitious technology deployment also alleviates demand destruction in aviation/shipping under stringent carbon constraints, enabling more service while meeting climate goals. The integrated framework highlights tradeoffs in fuel allocation (biofuels and electricity/H₂ competition across sectors) and upstream system responses (timing of electricity expansion and the mix of CCS/BECCS). Policymaking and targeted incentives (e.g., carbon pricing, fuel blending mandates, infrastructure investment) are pivotal to accelerate technology cost declines and deployment while managing intersectoral resource competition.

The study demonstrates that a zero-direct-emissions global transport sector by mid-century is achievable within a 1.5 °C-consistent pathway, contingent on rapid deployment of electrification for road and rail, substantial hydrogen adoption in shipping, and widespread use of sustainable aviation/shipping fuels—primarily FT biofuels, with e-fuels contributing more under high growth or constrained bioenergy. High transport ambition reduces total economy-wide sequestration needs and provides flexibility for industry and power sectors, while partially preserving transport service levels under stringent carbon constraints. The paper identifies mode-specific technology priorities and system-wide tradeoffs that can guide R&D, infrastructure planning, and policy. Future research should: assess specific policy packages and incentives; quantify supply chain, infrastructure, and critical mineral requirements; explore broader behavioral scenarios; evaluate additional fuel options (e.g., ammonia) and their planetary boundary implications; and integrate air quality and health co-benefits into decarbonization pathway assessments.

- Technology scope: Ammonia and some alternative fuels (e.g., DAC-to-methanol for petrochemicals) are not modeled; results may underrepresent roles of these options. - Biofuel pathway constraint: Aviation/shipping biofuels are limited to FT fuels from cellulosic biomass; alternative biofuel pathways and sustainability criteria could change availability, costs, and impacts. - Bioenergy cap and land-use: A global bioenergy limit (100 EJ by 2100; 70 EJ in sensitivity) is imposed to mitigate land competition, influencing BECCS deployment and sectoral fuel allocation; real-world land-use dynamics and policies may differ. - Cost/technology assumptions: Results hinge on GCAM v6.0 cost/performance inputs (e.g., DAC, electrolysis, hydrogen distribution, vehicle/vessel/aircraft technologies), which are uncertain, particularly for nascent aviation/shipping technologies. - Emissions accounting: Break-even carbon prices consider only direct emissions; upstream emissions for fuels depend on modeled decarbonization of electricity and hydrogen supply. - Demand behavior: Primary scenarios adopt SSP1 socioeconomics with reduced income elasticities and increased transit/ride-sharing; different behavioral trajectories can alter service levels and fuel needs (explored in sensitivities but still uncertain). - Model structure: Reduced-form climate (Hector), partial treatment of non-CO₂ GHGs (unconstrained), and representation differences versus other IAMs may affect comparability and absolute projections. - Mode aggregation limits: Walking/cycling omitted (minor shares); some within-mode heterogeneity (e.g., specific ship/aircraft classes) is abstracted in IAM representation.