This paper, A Generalized Method for Estimating Comprehensive Emissions for Electric Vehicles on a Per Trip Basis, presents a new methodology for estimating per trip GHG emissions for both EVs and equivalent ICEVs, and uses that methodology to produce estimates of total emissions for a representative sample of vehicles across 31,500 routes. Notably, our model takes into account both the well-to-wheel emissions of the fuel or electricity used by a vehicle, and important indirect sources of emissions such as vehicle manufacturing, maintenance, and disposal, which often aren’t included in vehicle emissions reporting standards. When accounted for comprehensively, the estimated emissions of all vehicles were found to be significantly higher than those given by more conventional models. Furthermore, indirect emissions from processes like vehicle manufacturing were found to be higher on average for EVs than for ICEVs. Finally, in regions that relied on fossil fuels for power generation, EV electricity emissions were correspondingly high, meaning that the potential emissions savings of EVs in these regions were lower than might otherwise be expected.
Both total EV emissions and their potential savings over ICEVs were found to be highly variable. Emissions from electricity generation were the most significant factor influencing variation in both total EV emissions and potential savings, with vehicle weight, battery capacity, and average driving speed also major contributing factors. The data also suggest that conventional models of emissions reporting may overestimate the potential savings of EVs by not including important sources of indirect emissions such as vehicle manufacturing and disposal, all of which tended to be higher for EVs than ICEVs. Despite this, EVs were still found to offer emissions reductions over ICEVs on average, and significant reductions in regions with well-developed clean energy infrastructure.
This study finds that both total EV emissions and potential savings over ICEVs are highly variable, and while EVs are less emissions intensive than ICEVs in many circumstances, the positive impacts of electrification heavily depends on ongoing investment and decarbonization efforts in other sectors, particularly power generation.
Overview
As the effects of anthropogenic climate change become more dire and the need to reduce greenhouse gas emissions more pressing, the development of practical emissions reduction solutions is becoming a higher priority for both policymakers and corporate actors. Within the transportation sector, one of the most popular proposed solutions is the replacement of conventional ICEVs (internal combustion engine vehicles) with EVs (electric vehicles). In 2019, the global transportation sector produced approximately 8.7 billion tons of direct carbon dioxide equivalent (CO2e) emissions, with around 74% of that originating from road vehicles that could theoretically be replaced with more environmentally friendly EVs without significantly disrupting operations. Unlike ICEVs, EVs do not burn fuel to operate, meaning that they don’t produce any direct or “tailpipe” emissions, which at first glance would make them seem like an unambiguously better alternative. However, the idea that EVs don’t produce any emissions is obviously false. The electricity EVs use is generated in power plants that often produce significant amounts of greenhouse gases. Furthermore, manufacturing, maintenance, disposal, and other aspects of the EV lifecycle also produce emissions, which are often higher than those for equivalent ICEVs.(1,2)
What are the emissions implications of EV adoption when all these often ignored factors are taken into account? In A Generalized Method for Estimating Comprehensive Emissions for Electric Vehicles on a Per Trip Basis we explore this topic in detail. We develop a more comprehensive model of EV and ICEV emissions, test the model using a representative sample of vehicles and routes, and analyze the results. Ultimately, we conclude that most other models tended to overestimate the potential emissions savings of EVs over ICEVs, as they did not account for important sources of indirect emissions which tended to be higher for EVs than for ICEVs. However, even when emissions were accounted for more comprehensively, we found that EVs still tended to be less emissions-intensive to operate than ICEVs on average. That said, our model’s estimates of EV and ICEV emissions varied significantly between different vehicles, route types, and especially regions, suggesting that use case is a very important factor in determining the potential emissions impacts of EV adoption.
Methodology
Those who wish to read a comprehensive description of the model used in this study should see Sections 3 and 4 of the full paper, which give detailed explanations of the methods used to estimate EV and ICEV emissions respectively. To summarize, we divide total emissions for both EVs and ICEVs into three major categories: embedded, operational, and end-of-life. Embedded emissions are the emissions from vehicle manufacturing and those produced when the vehicle is transported to the end user. Operational emissions are the emissions from the production, transport, and consumption of the fuel or electricity used by a vehicle, emissions relating to vehicle maintenance and repair, and emissions associated with the construction and maintenance of the road infrastructure the vehicle uses. Finally, end of life emissions are the emissions associated with the decommissioning and disposal of the vehicle and all its components. The sum of these categories of emissions is taken as the total emissions for a given route.
In order to assess this model, 15 routes were queried in each of 14 regions across the US and Europe in both summer and winter conditions for a set of 75 representative vehicles, for a total of 31,500 routes tested in all. A complete accounting of the results of these tests can be found in Section 6 of the full paper, with additional analysis of the model and its output in Section 5. A summary of some of the more pertinent results is given below. It should be noted that this model is deliberately pessimistic, using high or worst-case estimates for many aspects of EV lifecycle emissions, and therefore is likely to produce overestimates of total EV emissions and underestimates of emissions savings for many routes. Furthermore, the sample of vehicles was taken to be reflective of as broad a range of vehicles as possible, and is therefore not representative of average vehicle characteristics.
EV Emissions
Across all routes and vehicles tested, EVs were estimated to have produced around 320.2 gCO2e per kilometer driven on average. For the top five European economies by GDP this value was 302 gCO2e/km while for the top five US states by GDP it was around 300 gCO2e/km. However, this varied significantly depending on the vehicle, the route, and the region in which the route was calculated. On average, the largest single source of emissions for EVs was electricity (45.2% of the total), about a quarter of which was from electricity lost in various stages of the transmission and charging process (Figure 1). A further 18.5% of total emissions were from vehicle manufacturing, with slightly less than a third of these emissions deriving from the production of the EV’s battery on average. End of life and vehicle maintenance emissions were also major contributors to total emissions, while road infrastructure and distribution were relatively insignificant compared to the other categories. Routes driven during the winter were estimated to produce significantly more emissions than those driven in the summer (Figure 2). This is because EVs tend to operate less efficiently in cold temperatures, meaning that projected electricity consumption (and thus electricity emissions) were higher for winter routes. Notably, this means our results are likely overly pessimistic in warmer climates where the winter temperatures used in this study (-10 °C) are not representative.
There were notable differences in EV emissions between routes of different types (Figure 3), with rural routes having lower estimated emissions than suburban routes, which in turn had lower emissions than urban routes. The primary cause of this discrepancy was electricity emissions, which were higher for routes with lower average speeds. This is because EVs tend to operate less efficiently at lower speeds, meaning that they consume more electricity per kilometer driven and therefore have higher emissions.
Figure 4 shows average EV emissions by European country or US state. Every category of emissions except manufacturing and end of life varied somewhat between regions, but electricity emissions were by far the most important source of discrepancy. This was primarily because of differences in power generation infrastructure in between regions. In regions that have more developed clean energy infrastructure (e.g. Sweden and France), electricity is much less emissions intensive to produce on a per kWh basis than in regions that rely heavily on fossil fuels for power generation (e.g. West Virginia and Poland), which leads to correspondingly lower emissions.
Figure 5 shows average EV emissions by EU vehicle segment (3) as well as for cargo vans of different sizes (Small, Medium, Large, and Extra Large). Every category of emissions varied between vehicle classes, with manufacturing, maintenance, and end of life emissions increasing with battery capacity and all categories of emissions except electricity increasing with vehicle weight. Larger vehicles also tended to have higher electricity emissions than smaller vehicles due to their lower average electricity efficiency. Overall, smaller vehicles with lower capacity batteries tended to be much less emissions intensive to operate than larger vehicles with higher capacity batteries, as might be expected.
ICEV Emissions
Gasoline ICEV emissions were estimated to be higher on average than corresponding EVs, with a mean value of 437.4 gCO2e/km across all routes tested, 501.5 gCO2e/km for the top five European economies, and 354.3 gCO2e/km for the top five US state economies, while diesel vehicles were estimated to have produced 289.0 gCO2e/km on average across all routes tested, 306.71 gCO2e/km for the top five European economies, and 262.1 gCO2e/km for the top five US state economies. However, as with EVs, there was significant variance in emissions between route types, vehicles, and regions.
Across all routes tested, emissions from fuel were both higher overall and made up a higher proportion of total emissions for both gasoline (76.7%) and diesel (64.7%) ICEVS than electricity did for EVs (45.2%) on average. Emissions from all other categories, however, were both a smaller proportion of total emissions and lower overall for ICEVs than EVs (Figures 6 and 7). This was both because the production and consumption of fuel is much more emission intensive than electricity, which drove up fuel emissions, and because the ICEV lifecycle does not include the production, replacement, or disposal of large batteries, all of which are fairly emissions intensive processes, which drove down all other categories of emissions. ICEVs are also slightly lighter on average than EVs, leading to lower transportation and induced road maintenance emissions.
As with EVs, there was a strong correlation between average speed and total ICEV emissions, though the distribution varied between gasoline and diesel ICEVs. There were also significant differences in average ICEV emissions between vehicles of different classes (Figure 8). There were several factors that contributed to these differences, the most important of which was estimated fuel efficiency, which played a significantly more important role than did electricity efficiency in EVs. Vehicle weight was also an important factor, with increases in weight directly causing corresponding increases in manufacturing, distribution, maintenance, road infrastructure, and end of life emissions. Overall, smaller vehicles with less powerful engines tended to produce lower emissions, with diesel vehicles producing lower emissions than their gasoline counterparts, as might have been expected.
Vehicle Emissions Comparison
On average, EVs were estimated to have produced less emissions than their gasoline counterparts, with an average emissions savings of 117.2 gCO2e/km across all routes tested, 199.4 gCO2e/km for routes in the top five European countries by GDP, and 54.3 gCO2e/km in the top five US states by GDP. Conversely, EVs tended to produce more emissions than their diesel counterparts on many of the routes tested, with an average estimated emissions savings of -31.2 gCO2e/km across all routes tested, 4.7 gCO2e/km in the top five European countries by GDP, and -37.9 gCO2e/km in the top five US states by GDP. However, as might be expected given the nature of EV and ICEV emissions, potential savings varied significantly depending on the region, the vehicle, and route.
For all categories of emissions except electricity/fuel, ICEVs were estimated to have produced lower emissions than EVs (Figure 9). Therefore, the potential savings of EVs over ICEVs was found to be dependent on the emissions from the production of an EV’s electricity being significantly lower than the emissions from the production and consumption of the equivalent ICEV’s fuel. In regions with well developed clean energy infrastructure this discrepancy was found to be large enough to produce significant savings (for example EVs were estimated to have saved an average of 53.0% over gasoline ICEVs and 24.4% over diesels across all routes in France). However, in regions that heavily relied on coal or other fossil fuels for power generation, while EV electricity emissions were still lower than corresponding ICEV fuel emissions on average, the difference between them was not always sufficient to produce significant or even positive savings.
Estimated emissions savings were higher for rural routes than for suburban and urban routes (Figure 10), likely because EVs tend to operate significantly more efficiently at higher speeds, leading to lower electricity consumption and higher relative savings.
Savings also varied between regions, with five regions producing positive savings for both gasoline and diesel vehicles and two producing negative savings for both (Figure 11). This was due primarily to the emissions factor of electricity in these regions, with regions such as West Virginia and Poland (which rely heavily on coal power plants for electricity generation) producing lower savings than regions with more developed green energy infrastructure such as Sweden and France. Because the discrepancy in fuel/electricity emissions is the determining factor of emissions savings, regional differences in the emissions intensity of electricity was the primary cause of the significant variance in savings between regions. Finally, Figure 12 compares the total per kilometer emissions savings as produced by the model used in this study to the values produced by other, less comprehensive estimates. As can be seen, the other models produced much higher estimates of emissions savings between EVs and ICEVs. This is because these models do not account for embedded, end of life, or operational emissions except for fuel/electricity. Emissions in all these categories were higher for EVs than ICEVs on average, so excluding them tends to produce higher, and potentially overly optimistic, estimates of emissions savings.
Implications
The data from the current iteration of the model would seem to suggest that the emissions impacts of EVs are highly variable. The characteristics of a given EV, the type of ICEV it is replacing, the characteristics of the route, and especially the emissions factor of the electricity used can all lead to major differences in potential emissions savings between routes. Given this variance, the average values presented in this paper, while useful for understanding broad trends across the sample, are unlikely to be reflective of the emissions impacts of specific EVs. That said, while conventional vehicle emissions models may overstate the potential benefits of EVs, even when emissions were estimated using a less optimistic comprehensive model EVs were still found to produce emissions savings on average when compared to equivalent ICEVs, most notably in regions with well developed clean energy infrastructure. Furthermore, because electricity emissions were the primary factor in determining emissions savings, EVs have the potential to produce significantly greater emissions savings in the future as investment in clean energy infrastructure lowers the average emissions intensity of electricity. However, there is a more sobering side to this data. Even the best case emissions savings found in this study were still insufficient to meet the IPCCs 1.5° emissions reductions target for 2035, or its 2° target for 2050.(4) While our estimates of emissions savings were deliberately pessimistic, this would still seem to indicate that electrification of road vehicles may not be sufficient to meet our longer term emissions reductions goals by itself, and additional investments in both clean energy generation and less emissions intensive forms of overland transport may also be necessary to achieve our long term decarbonization goals.
End of abstract. Request a copy of the full paper below.
