Cross-cutting scenarios and strategies for designing decarbonization pathways in the transport sector toward carbon neutrality

China, the world’s largest CO2 emitter, has pledged to peak emissions by 2030 and achieve carbon neutrality by 2060 (“30/60”). Rapid growth in passenger and freight activity has made transport a major and rising source of emissions, accounting for about 10% of China’s total CO2 in 2019. Existing integrated assessment models often emphasize technological improvement (“Improve”) while urban and transport planning emphasizes demand and modal interventions (“Avoid” and “Shift”), creating a methodological gap. This study addresses the question: what long-term pathways and policy mixes can decarbonize China’s ground transport sector across regions by 2060? The purpose is to quantify the effectiveness, feasibility, costs, synergies, and trade-offs of policy measures within the Avoid-Shift-Improve (A-S-I) framework using an integrated, regionally resolved transport-energy model for 31 provinces over 2015–2060.

Prior global and national studies using integrated assessment models (IAMs) have quantified transport energy use and emissions, highlighting the mitigation potential of improving vehicle efficiency and decarbonizing fuels (e.g., electricity, biofuels, hydrogen). Transport planning literature, conversely, structures interventions around the Avoid-Shift-Improve approach, emphasizing land-use, compact development, mass transit, teleworking, and pricing. However, current IAMs typically have limited representation of land-use, infrastructure, and behavioral dimensions, while transport models often simplify energy system dynamics and cross-sector interactions. Previous work has explored electrification, hydrogen, and policy contributions to 2 °C and 1.5 °C goals, but few studies integrate demand reduction, modal shift, and technological improvements in a single, regionally detailed modeling framework for China. This paper fills that gap by combining a transport model with a detailed bottom-up energy system optimization model under an A-S-I scenario design.

The study develops a regional transport-energy integrated modeling framework covering China’s 31 provinces (mainland) for 2015–2060 at annual resolution. The framework couples: (1) a regional transport demand and mode choice model; and (2) a bottom-up energy system optimization model (AIM/Enduse) with detailed transport technologies. Transport model: Provincial passenger and freight demand are projected using panel data models with socioeconomic drivers (GDP per capita, population), land use (built-up area per capita), infrastructure (road length per capita), and generalized transport cost. Parameters are estimated from 2002–2015 provincial data using pooled, fixed-effects, and random-effects models with F and Hausman tests determining the preferred specification. Demand is split across seven ground modes: passenger (car, bus, two-wheeler, passenger rail) and freight (small truck, large truck, freight rail). Mode shares are determined via a cost-based discrete choice formulation using device, fuel, and time costs; parameters are calibrated to 2015 observed modal shares. Energy system model coupling: The transport model passes mode-specific service demands to AIM/Enduse, a recursive dynamic, linear optimization model that selects least-cost technology and energy mixes given a detailed database (technology lifetimes, capital and O&M costs, efficiencies, emission factors). AIM/Enduse outputs technology mix, costs, energy use, and emissions for transport, which are fed back to update generalized transport costs. Iteration proceeds until convergence in technology mix and costs. The integrated model is implemented in GAMS 33.2.0. Data and drivers: Historical socioeconomic, land, infrastructure, and transport data come from Chinese Statistical Yearbooks and sectoral sources; speeds, energy intensity, and load factors are from prior studies. Future GDP and population follow downscaled SSP2 to provinces; land use and infrastructure futures are based on national plans. Scenario design: A baseline business-as-usual (BaU) assumes continuation of current technological trends and existing policies. Twelve policy scenarios are constructed using the A-S-I matrix with four instruments each: Technology, Regulation, Information, and Price.

• Avoid: A_Tech (transit-oriented development via land-use intensity parameter convergence), A_Regu (compact city: 20% reduction in added built-up area by 2060), A_Info (teleworking calibrated from 2020 COVID-19 impacts), A_Pric (road pricing based on London-style congestion charging level by 2060).
• Shift: S_Tech (high-speed rail network expansion; mode preference convergence), S_Regu (bus priority; convergence to Shanghai’s preference), S_Info (carpooling/car-sharing: 50% higher load factors), S_Pric (fuel tax doubling gasoline/diesel prices by 2060).
• Improve: I_Tech (biofuel penetration: B50 by 2060), I_Regu (ICE bans with stringent EV/FCV penetration: 100% BEV for cars, buses, two-wheelers, small trucks; FCV for large trucks by 2060), I_Info (eco-driving and ITS: 5% efficiency gain), I_Pric (EV/FCV subsidies per China standards). Combined scenarios aggregate instruments within each strategy (Avoid, Shift, Improve) and aggregate strategies within each instrument (Technology, Regulation, Information, Price) to assess contributions and interactions.

• Under BaU, transport CO2 peaks at 1904 Mt in 2030 and declines to 1299 Mt by 2060 due to autonomous efficiency gains and demographic trends.
• All 12 A-S-I scenarios reduce emissions versus BaU. Peak emissions are lowered to 1844–1361 Mt depending on scenario (highest peak with I_Info; lowest with I_Regu). In 2060, I_Regu achieves the largest annual reduction, cutting 90% (1170 Mt) of BaU emissions, leaving ~129 Mt; A_Regu achieves the smallest reduction (~45 Mt in 2060).
• Cumulative 2015–2060 reductions: I_Regu is most effective (44% reduction); A_Regu is least (~3%). Avoid scenarios yield moderate cumulative reductions (3%–12%). Improve scenarios show the widest range (3% in I_Info to 44% in I_Regu).
• Combined strategies (2060 emissions vs BaU 1299 Mt): Avoid 862 Mt, Shift 466 Mt, Improve 72 Mt. Combined instruments: Technology 429 Mt, Regulation 44 Mt, Information 675 Mt, Price 756 Mt.
• Achieving a carbon-neutral pathway for ground transport via full A-S-I implementation yields an 81% cumulative reduction (59 Gt) in 2015–2060 CO2, with contributions (non-additive) of 28% (Avoid), 48% (Shift), 69% (Improve), and by instruments: 50% (Technology), 61% (Regulation), 36% (Information), 40% (Price).
• Drivers: Avoid reduces emissions primarily by lowering transport demand; Shift and Improve may increase demand but reduce energy and carbon intensity substantially, with Improve delivering the largest carbon intensity reductions via technology and fuel changes.
• Regional heterogeneity: Eastern/coastal provinces see larger reductions from Avoid, Information, and Price; southern/southwestern regions respond more to Shift, Technology, and Regulation; Improve shows a center-periphery pattern with larger effects in southwestern and coastal provinces. GDP per capita correlates positively with reduction potential for Avoid, Information, Price; negatively for Shift, Technology, Regulation; no significant correlation for Improve.
• Costs: Cumulative investment costs (2015–2060) decline most under Avoid (−97 trillion RMB) and Information (−95 trillion RMB). Improve and Regulation produce limited cost reductions (−6 and −3 trillion RMB). Generally, higher emission reduction potential is associated with higher investment, with Improve requiring substantial spending on EVs, FCVs, and ITS.
• Policy interactions: Improve can induce rebound travel and weaken Avoid’s demand reduction and Shift’s public transport promotion (e.g., EV subsidies can increase private car use). Synergies arise when Avoid (compact city) is paired with Shift (public transport) to mitigate congestion externalities, and when Improve is combined with Avoid/Shift to ease technology investment burdens.

The integrated regional transport-energy modeling demonstrates that China’s ground transport sector can follow deep decarbonization pathways when policy measures across Avoid, Shift, and Improve are combined thoughtfully. The analysis quantifies how different strategies affect demand, energy intensity, and carbon intensity, and highlights region-specific effectiveness and costs. It addresses the research question by showing that a balanced, regionally tailored A-S-I policy package maximizes synergies, minimizes trade-offs, and substantially contributes to China’s 2060 carbon neutrality objective. The findings bridge disciplinary gaps between IAM-style energy transitions and transport planning by explicitly modeling behavioral, infrastructure, and technological levers together, offering actionable guidance on which instruments to prioritize in which regions and how to combine them to achieve high reductions without excessive cost burdens.

This study contributes an integrated, province-level assessment of long-term decarbonization pathways for China’s ground transport, structured by the Avoid-Shift-Improve framework and enabled by coupling a transport model with a bottom-up energy system model. Results indicate that a comprehensive A-S-I policy package can reduce cumulative emissions by about 81% (59 Gt) from 2015 to 2060, with Improve delivering the largest carbon intensity reductions and Avoid/Shift shaping demand and modal structure. Policy recommendations include: co-planning compact, dense urban form with high-quality public transport; promoting clean vehicles through regulatory and financial incentives alongside ride-sharing and teleworking to amplify effects and mitigate costs; and designing region-specific packages reflecting heterogeneity in development stage, mobility patterns, and technology readiness. Future research should expand to include new transport technologies and modes (air and water), infrastructure requirements and costs, and behavioral economics to more completely capture system-wide impacts.

• Cost analysis included only device investment; behavioral costs of interventions (e.g., teleworking, modal shift) were excluded, potentially overstating net cost benefits of technology-focused scenarios and understating costs of behavioral scenarios.
• Infrastructure costs (e.g., EV charging networks) were not modeled, despite their necessity for decarbonization.
• Scope was limited to ground transport; air and water transport were excluded due to data limitations, though they contribute to overall transport emissions.
• Technological scope emphasized currently available or near-term technologies; emerging options (e.g., autonomous vehicles, personal aerial vehicles, delivery drones/robots) were not included.
• Potential effects of the COVID-19 pandemic on long-term economic growth and mobility were not incorporated into scenarios.