Exploring decarbonization pathways for USA passenger and freight mobility

The study investigates how the U.S. transportation sector—historically a persistent, large source of GHG emissions—could achieve deep decarbonization despite uncertainties surrounding technology, behavior, and policy. Against a backdrop of rapid advances in electric vehicles and clean fuels, evolving mobility options, and growing climate policy ambitions, the paper frames key questions: To what extent and how fast can U.S. passenger and freight mobility decarbonize? What combinations of technology adoption, policy measures, and behavioral change are required? The authors emphasize the sector’s prior characterization as difficult to decarbonize due to rising demand and slow fleet turnover, contrasting it with recent analyses and trends suggesting greater potential via electrification, low-carbon fuels (biofuels, hydrogen), efficiency improvements, and demand management. The purpose is to quantify pathways, sensitivities, and uncertainties in achieving substantial well-to-wheel GHG reductions by 2050, focusing on the roles of zero-emission vehicles (ZEVs), grid decarbonization, travel demand evolution, and sustainable fuels.

Previous economy-wide and sectoral assessments (e.g., IPCC AR5, U.S. Mid-Century Strategy, Stanford EMF 37) long suggested transportation would have substantial residual emissions in net-zero scenarios. More recent scenario work (IEA Net-Zero by 2050; U.S. Long-Term Strategy; Princeton Net-Zero America) projects deeper transport decarbonization through electrification, efficiency, mode shift, and low-carbon fuels, with EV shares rising rapidly by the 2030s–2040s. Yet skepticism persists regarding the sufficiency of individual levers: efficiency can counteract mode-shift benefits, electrification alone may not meet climate targets, and achieving large-scale behavior change (e.g., reducing car dependence) faces significant barriers. Evidence on shared mobility impacts is mixed, with some studies finding increased congestion and others showing emissions reductions in specific applications. There remains a paucity of high-resolution, integrated models capturing new technologies and nuanced mobility behaviors; TEMPO is positioned to fill this gap by representing heterogeneous consumers and mobility options at a national systems level.

The authors use NREL’s Transportation Energy & Mobility Pathway Options (TEMPO) system model to simulate U.S. domestic passenger and freight mobility across major modes (walking/biking, light-duty personal travel, Mobility-as-a-Service, public and intercity transit, domestic aviation and maritime, passenger and freight rail, and freight trucking). TEMPO endogenously models household-level travel decisions (demand, vehicle ownership, mode choice, technology adoption) across geographies and sociodemographics, and freight demand by operating segments, considering vehicle costs and fuel economy, fuel and electricity prices, charging/refueling access, trip characteristics, transit availability, and policies. The baseline scenario aligns with EIA AEO 2019 Reference, reflecting no significant progress in technology, behavior, or policy (e.g., low electrification, continued petroleum dominance), enabling consistent comparison. Study design comprises 2,173 simulations to 2050 with two core exercises: (1) univariate sensitivity analysis varying one input at a time (172 runs) to rank isolated impacts on well-to-wheel GHG emissions and ZEV adoption; (2) multivariate uncertainty analysis (2,000 runs) varying multiple inputs in combination to map a wide space of possible futures and identify deep decarbonization pathways. A total of 51 exogenous input categories (222 distinct assumptions) span vehicle technologies (costs, efficiencies), fuels and infrastructure (prices, charging/refueling availability), household behavior (ownership, trip frequency and length), system characteristics (transit availability, traffic/efficiency), and policies (ZEV sales mandates, carbon pricing). Multivariate combinations use Sobol low-discrepancy sampling with two grid cases: baseline and decarbonized electricity. Inputs are treated exogenously and independent; TEMPO endogenously captures mode/technology choices and feedbacks from cost, time, and availability changes. Uncertainties are larger in 2050 than near-term. The analysis focuses on well-to-wheel GHG emissions from mobility energy use (vehicle manufacturing and infrastructure life cycle are excluded). Biofuel options are not supply-constrained in TEMPO and aviation decarbonization is represented via biojet use only. No probabilities are assigned to inputs or outcomes.

• In isolation, the most impactful variable increasing 2050 mobility emissions is reduced on-road fuel economy: a 25% decrease raises emissions by 0.46 Gt CO2e (+29% vs. 2050 baseline). The most impactful single lever reducing emissions is a 100% light-duty ZEV sales mandate by 2030, lowering 2050 emissions by 0.44 Gt CO2e (−28% vs. 2050 baseline). For 2035 mandates: LD ZEV transition reduces 2050 passenger emissions by 0.35 Gt CO2e (−29% vs. passenger baseline) and MHD ZEV transition reduces freight emissions by 0.10 Gt CO2e (−26% vs. freight baseline). • Drivers of ZEV adoption (isolation): 100% ZEV sales mandates most strongly increase BEV stock. LD battery cost reductions to $40/kWh by 2050 yield ~45% LD BEV stock and reduce emissions by ~0.20 Gt CO2e vs. baseline. Charging convenience matters: zero value-of-time while charging pushes LD BEV stock to ~35% by 2050; each 2.9% increase in residential charging access raises 2050 BEV stock by ~1% (baseline home charging access 11%; aggressive case 75%). Energy price changes modestly influence LD BEVs; they are less impactful for MHD ZEVs if ZEV vehicle costs remain high. Cutting MHD BEV battery costs to $40/kWh drives ~41% MHD BEV stock by 2050. Light-duty FCEV adoption remains <1% under any single-variable change; MHD FCEV stock can reach ~540,000 with aggressive cost cuts but remains limited when varied in isolation without concurrent supportive factors. 100% household hydrogen access increases LD FCEV potential but still <1% stock without additional favorable cost/performance changes. • Multivariate uncertainty: Across 2,000 scenarios (1,000 per grid case), maximum 2050 decarbonization achieves an 89% reduction relative to 2019 (85% below the 2050 baseline). By 2030, maximum reductions reach 55% relative to 2019. In that 2030 best case: travel demand is 18% lower, vehicle occupancy 13% higher, aviation uses 1.7 billion gallons of biojet (≈13% of domestic aviation fuel), LD BEV stock is 35% (85% sales), MHD BEV stock 13% (69% sales), with grid emissions down 70% from 2019. • Electricity demand: 2050 direct battery-electricity demand for mobility has a median of ~1,000 TWh (range ~120–3,000 TWh) depending on ZEV adoption and travel demand; some scenarios see emissions rise if EV charging costs are high and demand and congestion increase. Affordable carbon-free electricity and charging access are critical. No deep decarbonization is achieved without a decarbonized grid; the mean emissions gap between baseline and decarbonized grids in 2050 is ~0.34 Gt CO2e (≈21% of 2050 baseline emissions). • Sensitivities within uncertainty: A 5% reduction in passenger trip frequency cuts 2050 emissions by a mean 0.042 Gt CO2e (baseline grid) and 0.039 Gt (decarbonized). Improving all-duty on-road fuel economy by 5% cuts 0.064 Gt (baseline grid) and 0.032 Gt (decarbonized). ZEV mandates are impactful in isolation but their average direct impact across uncertainty runs is moderated by competitiveness factors (vehicle/infrastructure costs, energy prices). • Deep decarbonization set (lowest 50 of 2,000 runs): mean 2050 emissions 0.27 Gt CO2e (83% below baseline 1.6 Gt; 87% below 2019). All occur under a fully decarbonized grid (BEVs, PHEVs on electric miles, and electrolytic-hydrogen FCEVs at zero grid WTW by 2035). Demand reductions are prevalent: 88% of deep cases reduce passenger or freight demand; 46% reduce both. LD ZEV stock averages 76% (96% sales) and MHD ZEV stock 68% (97% sales) by 2050; by 2035, LD sales average 82% and MHD 63%. Some pathways feature significant MHD FCEV shares (≈22% of deep cases) contingent on high hydrogen refueling access, low hydrogen prices (≤$6/kg by 2040), and improved FCEV costs/efficiency, often with longer shipment lengths and relatively pessimistic BEV assumptions. • Electricity for hydrogen: Indirect electricity for electrolytic hydrogen can reach up to ~3,100 TWh in 2050 (assuming ~51.3 kWh/kg H2). Deep scenarios with reduced demand average ~320 TWh indirect electricity for H2 in 2050; those with increased demand average ~750 TWh. Combining direct and indirect electricity, some scenarios exceed 4,000 TWh total transport electricity by 2050, especially with high FCEV adoption. • Biofuels: 2050 total mobility biofuel use across scenarios ranges ~1.5–61 billion gallons (0.12–7.8 EJ), with WTW emissions impact −0.092 to +0.25 Gt CO2e depending on fuel pathways and lifecycle intensities. Median 2050 biojet use ~8.7 billion gallons (≈51% of domestic aviation fuel). Deep decarbonization often entails moderate to high sustainable aviation fuel: without aviation demand reduction, ≈10–42 billion gallons/yr biojet may be needed by 2050 to deeply decarbonize domestic passenger travel.

The results indicate no single silver bullet; deep reductions require concurrent advances in technology, infrastructure, policy, and behavior. Rapid, widespread ZEV adoption is consistently central to decarbonizing on-road mobility, but outcomes vary widely with travel demand growth and system efficiencies. Demand reduction is a robust lever because it avoids emissions, reduces pressure on limited low-carbon energy supplies, and eases infrastructure requirements; however, it is challenging to implement due to behavioral, structural, and planning barriers. Even without demand reductions, deep decarbonization is achievable with very high ZEV sales by 2030–2040 and full grid decarbonization, but this implies large increases in electricity demand and careful system planning (including managed charging and flexibility). The relative roles of BEVs and FCEVs differ by segment and assumptions; MHD applications can feature meaningful FCEV adoption when hydrogen costs, refueling access, and vehicle performance improve substantially. Sustainable biofuels, especially for aviation, are critical to decarbonize modes that are difficult to electrify and to serve legacy fleets during transition. The interplay between the transport and power sectors is fundamental: mobility decarbonization depends on clean electricity, while electrified mobility can provide flexibility to support grid decarbonization. Managing these cross-sector dependencies and uncertainties is essential for planning resilient pathways.

Thousands of simulated futures show many feasible pathways to deeply decarbonize U.S. passenger and freight mobility by 2050. All deep decarbonization pathways include a decarbonized power sector and widespread ZEV adoption, supported by accessible charging/refueling, competitive clean energy prices, and improving ZEV costs and performance. Sustainable biofuels, particularly for aviation, complement electrification and are often necessary for deep reductions. Travel demand management substantially eases requirements for clean electricity and sustainable fuels. The study quantifies the range of transport-electricity needs (direct and indirect via hydrogen), highlights trade-offs between BEV- and FCEV-dominant futures, and underscores the importance of integrated transportation–power system planning. Future research should better capture endogenous demand shifts, long-distance travel alternatives (e.g., intercity rail, AV scenarios), infrastructure life-cycle impacts, and resource constraints (e.g., sustainable biomass) to refine decarbonization roadmaps.

The analysis focuses on well-to-wheel emissions from energy use and excludes vehicle manufacturing and infrastructure life-cycle emissions. Biofuel pathways are not supply-constrained or competed across sectors within TEMPO, potentially overestimating aviation decarbonization via biojet. Aviation technology alternatives beyond biojet are limited in the model. Changes in travel demand are largely exogenous; potential endogenous feedbacks (e.g., interactions between passenger and freight demand, cost-driven demand shifts, urbanization effects) are not fully represented. Autonomous vehicles and micromobility are not explicitly modeled due to data limitations, and detailed scenarios for intercity/high-speed rail buildout are not explored. The study does not evaluate hourly/seasonal grid dynamics or peak demands. Inputs and outcomes are possibilistic, with no probability assignments.