Introduction

To limit the probability of catastrophic climate change, greenhouse gas (GHG) emissions must decrease and approach net zero globally. However, annual emissions are still rising and have reached a new all-time high of 37.4 Gt CO2-eq. in 20231. On the positive side, the increase has slowed (from 2.1% in 2000–2010 to 1.3% in 2010–2020), and many countries have decreasing emissions, demonstrating that emissions and economic development are increasingly decoupled2. However, the time window for reaching climate goals without drastic measures is shrinking. Therefore, carbon budgets have been defined: To stay below a two-degree temperature increase with a probability of 50%, 580 Gt CO2 could be emitted globally as of 20183. At the current level of emissions and their trajectory, reaching this goal represents a huge challenge4.

Among the different sectors, transport is responsible for about 21% of global emissions4, of which road transport represents about 73%5. The primary decarbonization strategy for road-based transport is switching from internal combustion engines to battery-powered electric motors, which, combined with a cleaner electricity grid, will effectively reduce emissions. However, many conventional vehicles will continue to be sold at the same time, making further development of internal combustion engines (ICE) towards higher efficiencies an important pillar of the strategy6. The electrification transition has already begun and is supported in many countries by equivalent legislation7,8. The decline in the sales of electric vehicles in some regions is partly due to changes in government support, high interest rates, and a still lagging public charging infrastructure. These effects are, however, expected to be transitory9.

The highest ambition for supporting a swift transition to electric travel can be expected on the one hand in developed countries that have the required financial instruments, and in selected growing economies such as China and India, on the other, that benefit from having an own manufacturing industry10,11,12. Therefore, when assessing emissions and their relation to climate goals, one must consider the speed of the transition to electric vehicles (EV) and decarbonized electrical grids, and the development of the remaining fleet, together with their regional differences6. Additionally, alternative fuels can immediately reduce the emissions of the remaining fleet of ICEs not accessible for electrification (unless exchanged by a new electric vehicle), which could help to reach climate goals. Consequently, there is an ongoing debate about whether using alternative fuels in road transport makes sense, which is reflected in ongoing legislation in the EU13. The discussion revolves mainly around the higher efficiency of electric vehicles compared to ICE14,15 and the associated costs of EV and alternative fuels16,17,18,19,20. Simultaneously, using by-products from the aviation sector, which has introduced ambitious legislation and incentives to increase the share of alternative fuels until 2050, could become an option for road transportation. Many fuel production processes do not operate selectively but produce by-products in the gasoline and diesel range. The use of these fuels for defossilization is analyzed in this study as a complementary option to electrification.

Literature review

Previous studies have calculated the car sector’s emissions for different electrification strategies but do not vary the grid carbon intensity and neglect costs21,22,23,24, have a regional focus24, or do not account for emissions of the electricity used for charging vehicles25. Another study quantifies the costs and energy demand of reaching a net-zero goal in 2050 in the transport sector but does not consider different scenarios26.

In this study, we use three electrification scenarios to assess the well-to-wheels lifecycle emissions from the car transport sector and compare them to the sector’s carbon budget to reach climate goals. To this end, the scenarios comprise different sales shares of electric vehicles and carbon intensities of the grid electricity on a global level (including a break-down on a country level for the most important markets), which influence the composition of the fleet and the resulting energy demand and emissions. For these scenarios, we assess for the first time the possibility of reaching different climate goals using faster electrification, alternative fuels as by-products from aviation in a sector-coupling approach, and a successively reduced fleet size.

Methods

We estimate the emissions from the car sector until 2100 based on projected sales shares of electric vehicles (EVs) and cars with internal combustion engines (ICEs)27 (Fig. 1). Using data from Bloomberg28, we split global values into two groups of emerging and developed countries to capture their different timely adoption of EVs. We fit these projections to historical values for each country reported by the IEA to match current developments to the expected progression of sales shares. The fleet growth is based on data for OECD and non-OECD countries from EIA12, which we match to the categories of emerging and developed countries. We deduce the new vehicle registrations from the fleet size and specific car lifetime per country. Finally, we estimate vehicle emissions by considering car and battery manufacturing and consumption of fuel and electricity. In three scenarios (Business-as-usual (BAU), Middle-of-the-Road (MoR), and Optimistic), we vary the level of ambition regarding the transition to electrification and the decarbonization of grid electricity and analyze different options to meet climate goals by reducing the cumulative emissions of the car sector until the end of the century.

Fig. 1: Overview of the method.
figure 1

Projected EV sales shares in three scenarios by IEA (Stated policies, Sustainable Development Scenario, Net Zero by 2050) are split into developed and emerging countries with data from Bloomberg and fit to historical data on a country basis to represent different speeds of transition. The lifetime of vehicles in the fleet of each country is varied to find the most suitable value to represent the projected fleet size over time, enabling to identify the number of newly registered cars and hence the fleet composition in each country in every year for each scenario. Well-to-wheel emissions of EVs and ICEs are derived from the projected grid mix in each scenario combined with data for battery and vehicle production and assumptions about consumption of electricity and fuel.

Full size image

Regional resolution

We combine multiple data sources to address the need for regionally resolved forecasts of electric vehicle markets. This allows us to show the different technology adoption speeds, especially in contrast between developed and emerging countries.

IEA collects historical data (2001–2023) for EV sales, fleet, and new and total vehicle registrations for 31 countries (Supplementary Fig. 1). The remaining countries are aggregated as Rest of the World and Other Europe. These historical data are the starting point for our analysis.

Projection of car sales shares

We complement the historical car sales data by three scenario projections (Stated Policy, Sustainable Development, and Net Zero Emissions) for EVs published by the IEA27 and map them to our scenarios (BAU, MoR, Optimistic). These data reflect a global perspective without any distinction between countries or market types. However, the development of the electric vehicle market happens at different speeds in different regions: while developed regions act as early adopters with higher financial resources, electrification in emerging economies will follow at a slower speed. Historical data on electric vehicle car sales are available at the individual country level. Still, projections on the same level for all countries in our analysis and ideally from a single source are not available. To reflect different speeds of the transition to EVs in different markets, we therefore combine two datasets: global IEA projections divided into three scenarios for different speeds of transition and the projections of Bloomberg28, which have only a single scenario but are divided into an emerging, a global, and a developed country perspective. We do this by applying the relative differences between the “global” projection and “emerging” and “developed” from Bloomberg to the three EV sales share projections of IEA until 2050. We fit these six projections to the data’s historical development to obtain a forecast for each country or region from the underlying dataset. First, we categorize the countries into developed or emerging according to S&P BMI29. Second, we change the respective projections’ (developed or emerging) starting year until it matches each country’s last value from the historical dataset (Fig. 1 I bottom). To expand from the literature values, in the absence of reliable projections, we assume the sales shares to be constant from 2051 to 2100.

Projection of fleet size

We model the number of vehicles in the fleet for each country. EIA forecasts the vehicle stock development in OECD and non-OECD countries until 205012, which we match to the list of emerging and developed countries in our analysis. We use these growth rates to obtain the future number of vehicles per country, starting from the historical fleet size (Fig. 1 IIA). From 2051 to 2100, we extend the projections by EIA in the following way: for developed countries, the fleet size is assumed to be constant as the number of vehicles in this region is already projected to be near-constant by EIA; for emerging countries, EIA projects strong growth until 2050, which we then reduce by 2% each year until 2100 to reflect the increased level of wealth and the associated expected saturation of the car market.

Projection of new vehicle registrations

To apply the previously calculated sales share to the development of the fleets, we need to derive the number of newly registered vehicles, which are so far unknown. The newly registered vehicles consist of i) new vehicles entering the fleet and ii) replaced vehicles that leave the fleet at their end of life (Fig. 1 IIb). The former is the difference between the vehicle stock of two consecutive years. For the latter, we assume a Gaussian-distributed vehicle lifetime. In a similar analysis, Märtz et al.21 used a static global lifetime of 12 years per vehicle. This, however, leads to discrepancies between the cumulative fleet size (the sum of all newly registered vehicles over their lifetime) and the target fleet size by EIA12. Also, Held et al.30 found that vehicle lifetimes differ significantly between countries. To minimize the error, we adjust the vehicle lifetime individually per country in a recursive optimization loop to fit the EIA fleet size forecast12. The resulting vehicle lifetimes are close to the ones identified by Held et al.30.

While this method allows for the matching of cumulative new registrations with the designated development of the fleet size in a manner that matches literature values for vehicle lifetimes, it has several limitations: (i) no explicit import or export of vehicles between countries, (ii) our probability function for vehicle lifetime is a step function for each country that is smoothened with a Gaussian distribution over five years while Held et al.30 used a continuous probability density function, (iii) historical effects such as the drastic drop in new vehicle registrations during the Covid pandemic in 2020 are propagated through the analysis. While the latter creates a propagating effect throughout the analysis, which appears to repeat the pattern that led to its creation, this should not be interpreted as a projection of the exact cause (e.g., a global pandemic) repeating itself. It suggests that the historic decline in vehicle sales during the pandemic leads to a non-replacement and, therefore, lower vehicle sales after one lifetime of the vehicles. In other words, the cars not sold during the pandemic will not have to be replaced at the end of their lifetime. Of course, considerable uncertainty is associated with this projection, but this is an intrinsic uncertainty of future data projection. If the drop in sales during the pandemic would not appear in future sales in one way or another, this would be just another assumption, i.e., assuming that cars would be replaced earlier or later than their usual lifetime. Therefore, we accept the repetitive historical behavior resulting from the selected method and the historical data. Combining new vehicle registrations with electric vehicle sales shares gives each country’s fleet composition over time.

Extrapolation until 2100

The datasets and projections used do not extend beyond 2050. However, many ICEs will be in the fleet after 2050, which might impair the carbon budgets. To derive the cumulative emissions until 2100, the projections must be extended. Any underlying predictions are assumed to be constant after the year 2050, except for fleet growth in emerging countries, which will likely continue beyond 2050. We thus fix the EV sales shares and the carbon intensity of grid electricity at their respective values of 2050 and do not assume further development. For fleet growth until 2050, the projections of EIA12 for developed and emerging countries are used. From 2051 to 2100, the fleet size is kept constant at the 2050 value for developed countries. For emerging countries, on the other hand, the ten-year average fleet growth rate as projected by EIA from 2040 to 2050 is reduced by 2% annually and used to project growth until 2100, taking increasing wealth and market saturation of the market into account.

Vehicle emissions

LCA of EVs

The lifecycle emissions of an EV are comprised of vehicle and battery manufacturing, as well as grid electricity used to charge the battery. For the car without a battery, we assume 4.56 kg CO2-eq per kg of vehicle mass after Buberger et al.31, where this value is assumed to drop to zero until 2100 (assuming decarbonization of the economy). Each car is assumed to carry a 50-kWh battery. For battery manufacturing, 100 kg CO2-eq per kWh of capacity is taken from Kelly et al.32. We linearly reduce this value to zero in 2100 with the same reasoning as for vehicle manufacturing. We take the carbon intensity of grid electricity from 2020 to 2040 from IEA33 for the stated policies scenario (STEPS), the sustainable development scenario (SDS), and the net-zero scenario (NZE) and apply a break-down into countries and regions34,35. In our analysis, we map the STEPS scenario onto the BAU, SDS to MoR, and NZE to the Optimistic scenarios. We further assume an identical evolution of specific emissions in all countries. The consumption of electricity is assumed to be 15 kWh/100 km.

LCA of ICEs

For manufacturing an ICE vehicle, the emissions are assumed to be the same as for an EV without a battery. The combustion of one liter of gasoline emits 3.183 kg CO2-eq.21. We do not distinguish between gasoline and diesel vehicles but assume the same emission factor for a generic ICE vehicle. We assume a constant consumption of 5.9 L/100 km of ICE cars and an annual distance traveled of 15,000 km per vehicle (both for ICE and EV).

Selection of alternative fuels

We assume a selection of power-to-liquid fuels and different biofuels, i.e., HEFA fuels (hydrogenated esters and fatty acids) from the conversion of soybean oil and yellow grease, and Fischer-Tropsch (FT) conversion of herbaceous biomass, of municipal solid waste (MSW), of forestry waste, and agricultural residues. For HEFA fuels, the two indicated pathways serve as the basis for calculating costs and emissions, but the actual feedstock base is more extensive and comprises other waste oils, fats, and greases such as used cooking oil (UCO).

Fuel emissions

We use emissions for biofuels from Dray et al.18, who derive the values adjusting for future hydrogen and electricity emissions (assuming zero emissions for renewable electricity). For biofuels, the emissions of the fuel mix consisting of different fuel pathways are derived from the specific emissions of the single pathways18 and the breakdown of pathways in the potential analysis36. The biofuel supply is then assumed to always have this composition36, which is not a function of the scenarios used in our analysis. The emissions of power-to-liquid fuels are derived from the ICAO report on the feasibility of the long-term aspirational goal to defossilize aviation37, where the emissions for PtL derived from direct air capture of CO2 for IS1 are assumed for BAU, from IS2 for MoR, and IS3 for Optimistic. We do not distinguish between emissions of CO2 derived from waste sources or the air but assume all CO2 to come from the air.

Fuel availability

The available amounts of alternative fuels for the road transport sector are assumed to be the by-products of the alternative aviation fuels that often cannot be produced selectively, e.g., in the Fischer-Tropsch pathway. Several regulatory bodies, such as the United States and the European Union, have created incentives for producing sustainable aviation fuels (SAF). We assume a global scale-up of SAF shares in the aviation sector as defined by the ReFuelEU Aviation38 (2% in 2025 up to 70% in 2050). Furthermore, we assume that in 2100, 100% of the jet fuel will have to be SAF. From 2050 to 2100, we will use a linear scale-up of these shares from 70% to 100%. As SAF are often produced with by-products in the middle distillate range, we assume that 39% of the SAF fuel volumes appear as road fuels (gasoline, diesel)36. These shares are then multiplied by the total SAF volumes that are derived by assuming aviation fuel demand to grow by 1.1–2.8% per year (lower end of the range: 1.1% annual growth until 2050 and no further growth; upper end of the range: 2.8% yearly growth until 2100; Fig. 6). In each of the three scenarios (BAU, MoR, Optimistic), the same availability of alternative fuels is applied, assuming aviation to be defossilized globally as announced by ReFuelEU Aviation38 irrespectively of the road fuel sector.

We use the maximum fuel production ramp-up rates from Dray et al.18 for biofuels and PtL fuels that cannot be exceeded. When biofuels reach their assumed maximum potential of 190 Mt/y of road fuels, no further growth is allowed, even though some reports indicate that further biomass potentials exist via the growth of lignocellulosic plants on marginal lands39,40,41,42. For PtL fuels, we do not assume a maximum potential as there is no fundamental limit to their production (assuming CO2 from DAC). Unless limited by the maximum biofuel potential, we assume the total road fuels to comprise 50% of biofuels and 50% of PtL fuels. Once the maximum production potential of 190 Mt/y for biofuels is reached, only the PtL fuels are assumed to be able to scale up further.

Results

The transition to electric vehicles

The primary decarbonization strategy in the road transport sector is electrification43. Several countries and regions, such as the European Union, US states, Canada, and India, have announced a ban on internal combustion engines (ICE) until or before 204044. Combining EVs’ high efficiency with renewable electricity, the overall emissions of road-based transport could be brought to near zero (while already being locally emission-free). While many legislators recognize the long-term target, transitioning to electric vehicles requires a large-scale infrastructure adaptation, from establishing electricity generation, distribution, and transmission to local charging. This transition has already begun, but its completion will require time and large investments. Developing countries and remote regions will likely take more time in this process, and some sectors or regions may not be electrified at all. In this transitory phase, the fleet will comprise a growing share of electric vehicles and a remaining share of cars with internal combustion engines.

As the electrical grids are being decarbonized, the emissions of electric vehicles are also decreasing (towards zero for a fully renewable chain). In contrast, the emissions of ICEs depend on the fuel used. The emission reduction potential for fossil fuels is about 17% when all process steps in the supply chain are decarbonized45 while alternative fuels can decrease the emissions by 80–90% or even more46,47,48,49. The emissions from passenger cars and light-duty vehicles (LDV), therefore, depend on the speed of this transition (involving the generation of renewable electricity, electrical grids, EVs, and charging stations) and to which degree alternative fuels are used.

Figure 1 analyzes the transition to EVs based on renewable electricity availability and EV sales development in different regions. Three cases are defined: (i) a business-as-usual (BAU) scenario, (ii) a middle-of-the-road (MoR) scenario, and (iii) an optimistic scenario. The scenarios’ sales shares50 and grid electricity emissions are based on the IEA scenarios for stated policies, sustainable development (compliant with the Paris climate goals), and the net-zero goal (limiting global warming to 1.5 °C), respectively. Additionally, we distinguish between developed and emerging countries29 because of their fundamentally differing underlying developments. While the markets in developed countries such as the US or most of the EU are saturated concerning the total number of vehicles, strongly growing economies such as China or India will experience significant growth in the vehicle fleet.

In developed countries, the sales share of EVs is projected to increase faster, mainly due to the higher available financial resources (both on a national and personal level) that allow to compensate for the higher cost of electric vehicles today. In the BAU scenario, the maximum EV sales share only reaches about 40%, leaving most of the future fleet composed of ICEs. In the MoR and optimistic cases, the sales shares of electric vehicles reach 80% or more, converting the fleet to mostly EVs by 2040.

The resulting emissions of the ICE fleet are already falling today, except for the BAU scenario in emerging countries (which see a significant increase and a plateau towards 2100), where the increasing number of ICE cars leads to still rising emissions. The emissions of EVs experience a peak in the 2030 s for developed countries and around 2040 for emerging countries. This behavior is explained by the quickly growing fleet of EVs on the one hand and the decreasing carbon intensity of the electricity grids (and production of vehicles and batteries) on the other. The emissions keep rising with the growing fleet of EVs until grid decarbonization overturns this dynamic. As this process happens faster in developed countries, the peak in EV emissions is seen in the 2030 s. Due to slower grid decarbonization, emerging countries peak in the 2040 s. The larger speed of electrification is also the reason why EV emissions are higher in the MoR (in emerging countries) or Optimistic scenarios (in developed countries) than in the BAU scenario. The different dynamics in electrification and grid decarbonization lead to the different order of scenarios in these two regions. In all scenarios and regions, the total emissions decrease from the BAU to the MoR and to the Optimistic scenario (see SI). These results agree with the findings of the IPCC AR6 WG III (Chapter 10)51, assuming similar emission reductions for the car and transport sectors (see SI for more information).

The scale of the problem

Most of the cumulative emissions are attributed to the ICEs in the fleet (remaining vehicles and newly registered ones, see Fig. 2). Figure 3 shows the cumulative emissions until the end of the century for the three scenarios broken down into vehicle type and time frame (today-2050 and 2051–2100). The emissions of the BAU scenario significantly exceed the carbon budget for all climate goals (1.5, 1.7, 2.0 °C): close to 300 Gt CO2-eq. are emitted until 2100, about two-thirds of which in the second half of the century. In this case, the largest share of emissions is due to the ICE fleet, which is decreasing in developed countries but increasing (although the respective sales share is decreasing) in emerging countries (see Supplementary Fig. 5 for more information). The yearly emissions drop insignificantly until the end of the century because the fleet increase in emerging countries compensates for the drop in emissions by electric vehicles, as can be seen clearly in the example of China (see emissions breakdown in Supplementary Fig. 5). In the MoR and Optimistic scenarios, the cumulative emissions stay below the two-degree budget but above the limit for a 1.5 °C-increase. The difference between the MoR and Optimistic scenarios is the latter’s faster and more extensive fleet electrification. This leads to a smaller ICE fleet and, thus, lower associated emissions, especially after 2050. If the ambitious electrification rates of these two scenarios can be reached, the Paris climate goal is achievable in the road-based transport sector. If, however, electrification should take a pace as in the BAU scenario or even slower, the climate goals are out of reach, and additional measures must be taken.

Fig. 2: Modeled sales shares, fleet composition, and emissions for EVs and ICEs for three scenarios based on projections by the IEA12 and Bloomberg28 (see methods).
figure 2

The business-as-usual (BAU) scenario corresponds to current policies. In contrast, the Middle-of-the-Road and Optimistic scenarios are based on the EV sales shares and electricity grid decarbonization compliant with the Paris climate goals and 1.5-degree warming, respectively. The results further distinguish between developed countries and emerging countries to highlight their differing behavior regarding the uptake of EVs and fleet composition. Fast electrification of new vehicles is a very effective measure to reduce the emissions of the fleet in combination with rising shares of renewable energy in the electrical grids. The majority of the global fleet will be found in emerging countries by the middle of the century.

Full size image
Fig. 3: Cumulative emissions from EVs and ICEs in different scenarios until 2100 (breakdown into various countries in SI).
figure 3

The horizontal lines indicate carbon budgets61,62 (62 Gt, 142 Gt, 261 Gt) for the car transport sector to reach the climate goals with a warming of 1.5, 1.7, or 2.0 °C, respectively, until the end of the century. The Middle-of-the-Road and Optimistic scenarios reach a warming of 1.7 °C or less by a swift transition to electric vehicles, thus avoiding emissions from a large fleet of ICEs in the second half of the century.

Full size image

How do we reach the climate goals?

The road-based transport sector emits significant amounts of GHGs, which are not compliant with the climate goals in the business-as-usual case. A more considerable effort, including a diverse set of measures, may be required to limit emissions to a lower level in these cases. For each of the three scenarios, we analyze the effect of three measures to reach the climate goals (2.0, 1.7, 1.5 °C global warming until the end of the century).

Speed up current development—exclusively new electric cars

To avoid excessive emissions, a large ICE fleet using fossil fuels must be avoided, especially in the second half of the century. Thus, as a first measure, faster electrification is proposed—an obvious choice based on the projected emissions in the different scenarios (see emissions from ICEs and EVs in Fig. 3). We assume that from a given year on, only electric cars can be registered (Fig. 4). This goal has been set for 2035 in different regions of the world, i.e., the European Union and California, via a ban on conventional ICEs. Reaching this goal globally in 2035 would decrease the emissions significantly, especially in the BAU scenario, leading to lower cumulative emissions by about 270 Gt CO2-eq., until 2100. In the MoR and Optimistic scenarios, the reduction is still significant at about 50 Gt CO2-eq. These saved emissions are equivalent to moving the car sector by 0.7 °C or 0.1 °C closer (for the BAU or MoR/Optimistic scenarios) to its respective sector-related climate goals.

Fig. 4: Cumulative emissions from the car transport sector in three scenarios under the assumption of full global electrification (global EV sales share 100%) in different years (see SI for more information).
figure 4

The Paris climate goal is reached irrespectively of the year of full electrification. While the reference cases (Fig. 1) only assume full electrification in the most ambitious case (developed countries, optimistic scenario), the measure shows great potential by avoiding an ICE fleet in the second half of the century.

Full size image

We find that the specific year of an ICE ban has only a minor influence on reaching the climate goals: even an ICE ban in 2050 would result in corresponding sector-specific emissions compatible with the Paris climate goals in all scenarios. Of course, an ICE ban presents a challenge, especially in emerging countries, where the electrical grid may not be adapted to the distributed loads, and there is a lack of charging infrastructure (including developing local grids). Likewise, countries with a distributed population may need help reaching full electrification soon due to the widespread infrastructure that must be implemented and EVs being most efficient in an urban environment. Nevertheless, it represents an effective instrument for reducing emissions, especially in the BAU scenario.

Limit car use globally—reduced fleet growth

The second proposed measure to reduce the emissions is a reduction of the fleet growth rate as projected by EIA12, which is used as a standard in all scenarios (Figs. 2, 4). In the EIA projection, the global car fleet is projected to grow from 1.4 billion today to about 2.2 billion in 205052. Significant growth is only expected in emerging countries, while the fleet size in developed countries is almost constant (Fig. 2). As a measure to reduce emissions, here we reduce the projected global growth by 0–100% and calculate cumulative emissions (Fig. 5). By continuing the projection, assuming a constant fleet size in developed countries after 2050 and gradually decreasing growth in emerging countries (by 2% annually, see), the global car fleet will reach 2.5 billion units at the end of the century. Limiting this growth rate reduces the number of vehicles without changing the projected electrification rate. This effectively cuts the number of newly registered cars, mainly affecting emerging countries with slower electrification rates. Consequently, the number of new ICEs entering the fleet is reduced, reducing the emissions. This effect is most significant in the BAU scenario, where the EV sales share reaches only 40% in emerging countries.

Fig. 5: Total cumulative emissions from the car transport sector until the end of the century for a reduced growth rate relative to the EIA projection.
figure 5

Most of the total growth is expected in emerging countries, while the fleet remains almost constant in developed countries. As the EV sales share in emerging countries is projected to be lower than in developed countries, this measure would effectively prevent the sales of mostly ICEs and therefore reduce emissions. A reduced growth rate would, mainly impact emerging economies and is therefore seen as critical.

Full size image

Use synergies in the system – alternative fuels as by-products from the aviation sector

Reducing emissions in the car sector does not necessarily require electrification. It can also be achieved by using alternative fuels with lower specific emissions than fossil fuels. Currently, these fuels are used in the car sector as either ethanol and biodiesel (mixed with gasoline and diesel) or in the aviation sector, where they are produced from plant oils and represent a core pillar of the defossilization strategy53. A significant volume increase in ethanol and biodiesel production would cause competition for land with food and fodder production and is likely of limited potential. As legislation in the EU and other regions requires rising shares of alternative fuels in aviation, the production volumes are expected to increase over the following decades. Various fuel pathways are being developed, using biological and non-biological carbon as a resource base to satisfy the demand. Most of these have in common that they are not selective, i.e., they not only produce jet fuel but a mix of different hydrocarbons, including gasoline54. It is likely that large amounts of gasoline as a by-product will become available from the aviation sector, which could help to reduce emissions in the automotive industry. To estimate these by-products’ amount, we assume that the required shares of alternative fuels in ReFuelEU Aviation38 (2% in 2025, rising to 70% in 2050) are reached globally, following the emission reduction goal of net-zero emissions by 2050 as suggested by IATA53. We use a fixed by-product share of 39% of road fuels concerning aviation fuel volumes36. Furthermore, aviation fuel burn is assumed to grow annually by 1.1–2.8%18, with growth estimates subject to large uncertainties. To cover a broad range of fuel demand in the aviation sector, we use a lower bound of 1.1% fuel demand growth until 2050 (and constant fuel demand after that) and an upper bound of 2.8% annual growth until 2100. The resulting by-product volumes are shown in Fig. 6, where we explicitly contrast the expected fuel production in these aviation growth scenarios with the theoretical maximum fuel production capacity18. While we do not further disaggregate techno-economic feasibility or policy risk, the upper and lower aviation growth bounds serve as a proxy for uncertainties in supply scale-up. We acknowledge that additional constraints—such as fuel selectivity, land use, and lifecycle emissions—could influence actual supply and identify these as areas for future analysis. A detailed treatment of land use and lifecycle emissions is beyond the scope of this paper. In the MoR and Optimistic scenarios, aviation by-products could cover the demand of the remaining ICE fleet in the 2050ies even for a slowly growing aviation sector. For a strong growth of aviation, even the remaining fuel demand in the BAU scenario could be covered towards the end of the century, where the expected fuel production from aviation does not exceed the capacities of expected maximum ramp-up rates of fuel production facilities18.

Fig. 6: Projected road fuel volumes as by-products from the ramp-up of alternative fuels in the aviation sector (shaded area).
figure 6

For aviation, annual growth rates of 1.1% (until 2050) to 2.8% (until 2100)18 and alternative fuel shares according to ReFuelEU Aviation38 (starting at 2% in 2025 and reaching 70% in 2050) are assumed. Based on the calculated vehicle fleets, the expected fuel demand in three scenarios (BAU: brown, MoR: green, and Optimistic: yellow lines) and the maximum fuel production rate18 (dotted line) are also shown (assuming a fast scale-up of alternative fuel production). In the MoR and Optimistic scenarios, alternative fuels as by-products from aviation could cover the remaining fuel demand of the ICE fleet even for slow growth of aviation. In contrast, only a strong growth of aviation could enable this in the BAU scenario towards the end of the century.

Full size image

Assuming all the liquid by-products from aviation become available for road transport, significant emission reductions are achievable: in the BAU scenario, cumulative ICE emissions could decrease by 12–28% (49–115 Gt), by 23–42% (32–58 Gt) in the MoR, and by 31–33% (34–37 Gt) in the Optimistic scenario (Fig. 7). The overall impact on emissions is limited because of the time it takes to ramp up fuel production and the road-based transport sector’s larger fuel demand than the aviation sector’s supply. Assuming strong growth of the aviation sector (see SI for more information), the annual emissions of the car sector can be reduced by more than half towards the end of the century, irrespective of the scenario. Considering that alternative fuels will appear as by-products from this sector, their contribution towards reducing emissions in the car sector can thus be precious. A further reduction of ICE emissions would require larger volumes of by-products from, e.g., the shipping sector or through dedicated production (in the absence of demand-side measures). In this case, the maximum foreseeable output (Fig. 6, gray dotted) would cover the remaining fuel demand of the car sector in the 2040ies (for the MoR and Optimistic scenarios) or towards 2060 in the BAU scenario. Of course, the actual alternative fuel share will be determined by costs, which, in detail, exceeds the scope of this study.

Fig. 7: Emission reductions in the road transport sector by using alternative road fuels as by-products from the aviation sector in three scenarios (assuming a global adoption of ReFuelEU Aviation38 introducing alternative fuels).
figure 7

The uncertainty is due to different assumed annual growth rates of aviation (between 1.1% until 2050 and 2.8% until 2100). The emissions savings are significant and help reaching the Paris Climate Goal in every scenario. A total between 32 Gt (Optimistic Scenario at slow growth of the aviation sector) up to 100 Gt of CO2 (in the BAU scenario for strong growth of the aviation sector) could be saved by supplying alternative fuels to the remaining fleet of ICE cars.

Full size image

Discussion

For many people, personal road transport is an essential part of daily life and enables them to travel between families and communities or commute to work. It is also a significant source of global emissions, threatening to overburden the global ecosystem. Bringing our emissions to zero and maintaining our standard of living is one of the major challenges we face in this century. Accomplishing this task in the transport sector requires a global effort by switching to a renewable energy base. Considering the urgency and magnitude of the efforts and investments involved, a clear strategy for decarbonizing the car sector is required.

With the analysis above, we aim to quantify the effects of different pathways towards reducing emissions in the car sector. Below, the most important findings are summarized and discussed.

Emissions from the car sector represent a significant share of the total emissions and must be reduced to comply with climate goals

About 9.5% of global CO2 emissions are emitted by cars today55,56. If only stated policies are enacted, the cumulative emissions until 2100 will surpass the car sector’s share of the two-degree carbon budget. Additional efforts have thus to be made to reduce the emissions.

Emerging countries experience substantial growth in their car fleets, while developed countries have an almost constant number of vehicles

This is mainly due to the correlation between GDP and car ownership57, which leads to fleet growth occurring primarily in the large emerging economies of China, India, and Brazil. Therefore, an effective reduction of future emissions is tied to a strategy addressing both new registrations and existing cars in the fleet. A global effort to enable a quick transition to an electric vehicle infrastructure especially in the emerging economies could help to prevent the sale of a large number of ICE vehicles using conventional fuels.

Electrification is the most effective means of reducing new car’s emissions while maintaining equitable mobility

In most economies, EVs already achieve lower emissions than ICEs today. If electrification cannot be scaled up quickly, the car fleet will be composed of ICEs for a long time. To comply with the climate goals, these vehicles must be defossilized with alternative fuels. Otherwise, more drastic measures must be taken to limit global car sales or use. Because global fleet growth mainly happens in emerging economies, limiting car sales would hinder these economies’ growth. Considering that developed countries have contributed the most emissions to this day starting from the industrial revolution and the establishment of economies built on the conversion of coal and oil, this measure would be questionable to achieve an equitable distribution of efforts to prevent climate change. A more promising approach would be the establishment of an EV infrastructure in strongly growing countries to prevent the global growth of the ICE fleet. This could potentially happen via bilateral or global efforts including but not limited to a global emissions trading system that facilitates the financing of measures to reduce the emission of greenhouse gases where it is the most economical.

Alternative fuels are a suitable way to reduce the emissions of the remaining ICE fleet and are likely to be available in sufficient quantities as by-products from other sectors such as aviation

Alternative fuels are currently under development and can reduce specific emissions by more than 90%47. Sectors such as aviation and shipping depend on the continuous use of liquid fuels, which, together with regulatory decisions, will create a large market for alternative fuels. The production of these fuels is, in many pathways, not selective: by-products will be produced that could be used in the road transport sector to decrease the emissions of the remaining ICE fleet. For large-scale production, the costs of alternative fuels could drop to values below the market price of fossil fuels in most countries18,38,58,59,60. To establish a quick ramp-up of these production volumes, a global effort is required creating an (ideally unified) regulatory framework that allows to export alternative fuels to other markets and the necessary ecosystem to attract investments. To this end, a reliable and ambitious legislation ramping up the shares of alternative fuels and the creation of financial instruments to attract long-term investors needed to support the scale-up of the technology will be crucial.

The most effective strategy to reduce the car sector’s overall emissions relies on quick electrification and defossilization of the remaining ICE fleet with alternative fuels

As a global option for equitable and climate-friendly mobility, electrification requires renewable electricity and a large-scale charging infrastructure. This process will likely require decades to complete, and in the transitory phase, ICEs comprise a large fraction of the car fleet. The only way to defossilize these vehicles is by exchanging the fossil fuel by a renewable one. Alternative fuels can effectively reduce the emissions of the remaining fleet of ICEs at potentially competitive cost if produced at scale and if synergies with other sectors (aviation, shipping) are leveraged. With this strategy, new electric vehicles and the remaining ICEs can be addressed, and the automotive sector’s total emissions can be reduced most effectively. To this end, regulation should address this challenge with a technologically open but ambitious framework that allows the defossilization of vehicles with either electrification or alternative fuels and that accounts reliably and transparently the emissions created in either case. This will open a long-term perspective for investors required to finance the scale-up and establishment of infrastructures of both technologies. In the end, there is no need to be on the fence.