Strategic roadmap for optimising vehicle emission reductions and electrification

Prompted by policy support, battery electric vehicles (BEVs) have become increasingly popular in many countries and economies. To ensure that vehicle electrification contributes to reduction in emissions, governments should develop appropriate transition plans that consider the lifecycle CO2 emissions of these vehicles. In this study, we aimed to establish an emission reduction-focused transition trajectory for vehicle electrification using lifecycle optimisation. Through a Japan-centric case study spanning from 2005 to 2055, we identified an optimal fuel-type progression for car owners, underlining the potential for BEVs to be introduced in the 2030s, a decade ahead of the baseline, if higher emission reduction can be attained. Policymakers are advised to facilitate a gradual shift toward hybrid electric vehicles and plug-in hybrid electric vehicles that initially outperform BEVs in emissions, until a robust level of lifecycle CO2 reduction is achieved within the automotive sector. This study contributes to the discourse by offering a strategic roadmap for maximising emission reduction through targeted vehicle electrification, making it pertinent and informative for both policymakers and stakeholders. The insights underscore the critical role of deliberate policy interventions in orchestrating a sustainable and effective transition toward a lower-emission transportation paradigm.


Introduction
Electric vehicles (EVs) are gaining increasing acceptance owing to policy support in many countries and economies and have surpassed 26 million units worldwide in 2022, counting over five times more than in 2018 [1].However, the mitigation effects of EVs are currently marginal in regions that are highly dependent on fossil fuels [2][3][4][5].This is because the emissions from electricity generation for EVs (i.e.well-to-tank emissions) depend on the energy dynamics of the region in which they are charged [6].In regions with low/decarbonisation of the power sector, the lifecycle CO 2 (LC-CO 2 ) emissions of EVs are lower than those of conventional vehicles [7,8].
For vehicle electrification to contribute to emission reduction, it is essential to reduce their LC-CO 2 emissions, which entails decarbonising the power sector and improving emissions from the vehicle and battery manufacturing phases, in addition to improving energy efficiency in the driving phase [9].Whereas there remain regions where this is not yet feasible, and given the various technological innovations in the product lifecycle, policymakers should begin to more actively promote EVs to reduce emissions [6].However, the roadmap for vehicle electrification remains uncertain with regard to the optimal timing for policymakers to promote EVs to ensure their effective contribution to emission reduction, taking into account the future energy mix and technological innovations.
Lifecycle optimisation (LCO) provides optimal replacement strategies for a specific time period, with the aim of minimising the environmental impact and costs of the target product over its entire lifecycle [10].Despite improvements in technological efficiency, studies have shown that the prolonged use of older products (e.g.laptops [11], household refrigerator [12], mid-sized saloons [13]) results in lower LC-CO 2 emissions and costs than replacement with newer, more energy-efficient products.Although blanket vehicle electrification may not necessarily lead to emission reductions [2][3][4][5], the optimal replacement pathway for transitioning to EVs to achieve effective emission reduction has not yet been identified.Therefore, based on LCO, we have derived a transition strategy that maximises the contribution of vehicle electrification to emission reduction.Subsequently, we examined whether the results for EVs, as discontinuous technological innovations, align with the findings of studies that suggest long-term product ownership is desirable [11][12][13].We also determined the levels of technological innovation and emission reduction that are required for EVs to contribute to climate mitigation.
Based on the environmental input-output lifecycle assessment [3,9], we estimated the LC-CO 2 of five fuel types of passenger cars: gasoline vehicles (GVs), hybrid electric vehicles (HEVs), plugin hybrid electric vehicles (PHEVs), battery electric vehicles (BEVs), and fuel cell electric vehicles (FCEVs).Additionally, based on Tokito et al [14], we identified an optimal replacement strategy for a car owner by modelling the shortest path problem [15].Furthermore, by finding the top-k shortest paths using the k-best algorithms [16,17] and comparing multiple solutions, we have designed robust and feasible strategies.By integrating life-cycle assessments with this methodological approach, we can improve the existing LCO model by (1) calculating resilient transition strategies that consider changes in environmental factors (e.g.energy mix) and technological innovation for cars and (2) exploring the disparities between the ideal pathway and car owners' choices seeking to minimise ownership costs.
We selected LC-CO 2 and the ownership cost as objective functions that were minimised in the shortest path problem.Our approach successfully achieved an optimal strategy comparable to that identified by other methods [10,13,18], validating the effectiveness of the replacement strategy.Based on our findings, we propose optimal transition policies to maximise the contribution of vehicle electrification to decarbonisation.

Lifecycle CO 2 emission
To calculate LC-CO 2 , we assumed that the system boundary of automobiles consists of direct and indirect emissions in the following eight life stages: (1) manufacturing; (2) refinement of gasoline and/or power generation consumed (well-totank); (3) burning of gasoline by GVs, HEVs, and PHEVs (tank-to-wheel); (4) repair and maintenance; (5) installing an EV home charger (first-time EV purchase only); (6) battery replacement (except for GVs); (7) car inspection, testing for conformity with many maintenance regulations; and (8) scrapping of endof-life cars.Based on an environmentally extended input-output analysis [3], we estimated the LC-CO 2 emissions for each life stage (excluding the tank-towheel phase; see equations (S1)-(S3).The inputoutput table in Japan used in this study, the inputoutput table for analysis of next-generation energy systems, contains domestic monetary transactions for 153 products [19,20].We collected data on the input requirements for each life stage for vehicles using each fuel type from Washizu and Nakano [19] and Nakamoto et al [3] (tables S1 and S2).
The LC-CO 2 , Q (n, i, j) for a car manufactured and purchased in year n and owned for i years with fuel type j can be formulated as follows: where Q mfg (n, j) , Q chg (n, j) , and Q dis (n, j) denote the CO 2 emissions during the pre-consumer, EV home charger installation, and scrapping phases for a car manufactured in year n with fuel-type j.Q w2t (n, i, j) , Q t2w (n, i, j) , Q r&m (n, i, j) , Q ins (n, i, j), and Q bat (n, i, j) represent the CO 2 emissions over i years during the well-to-tank, tank-to-wheel, repair and maintenance, car inspection, and battery replacement phases for a car of fuel-type j manufactured in year n, respectively.For any manufacturing year n, Q (n, 0, j) represents the scenario in which no new car is purchased during that year.

Ownership cost
Vehicle operating costs by fuel type include the following items: vehicle cost (PHEV, BEV, and FCEV include subsidies), fuel and/or electricity purchases during the driving phase, repair and maintenance, EV home charger installation, vehicle inspection, battery replacement, disposal, taxes (purchase tax, weight tax, vehicle tax, and liquid fuel tax), and resale value (assuming a 10% reduction in vehicle cost per year; table 1).Each cost is estimated by referencing to a specific representative model developed for each fuel type.For passenger cars, the Japanese vehicle inspection system applies to vehicles three years after new purchases and every 2 years thereafter [21].We assumed that the cost of ownership depends on the manufacturing year, fuel efficiency, and other factors.Taxes and subsidies were calculated by applying the tax and subsidy programmes as of April 2023 to vehicle specifications by fuel type.The exchange rate between the U.S. dollar and the Japanese yen was assumed to be USD 1 = JPY 131.57[22].The ownership cost,C (n, i, j) C (n, i, j), for a car manufactured and purchased in year n and owned for i years with fuel-type j can be formulated as follows: where C mfg (n, j) , C chg (n, j) , and C dis (n, j) denote the cost for the pre-consumer, EV home charger installation, and scrapping phases for a car manufactured in year n with fuel-type j, respectively.
), and C bat (n, i, j) represent the costs over i years during the driving, repair and maintenance, car inspection, and battery replacement phases for a car of fuel-type j manufactured in year n, respectively.For a given manufacturing year n, C (n, 0, j) represents the scenario in which no new car is purchased during that year.

Model construction
A decision maker who chooses to replace or keep the product in each period has 2 N possible combinations of actions in period N (n = 1, 2, . . ., N) if only one product choice exists.Furthermore, if there are multiple product choices (five in this study), numerous possible combinations of actions arise in period N (6 N in this study).In the LCO model [10,18], an optimisation problem is set up to efficiently identify the optimal path from several path combinations.We defined the fuel-type variables as follows: where t n ∈ {GV, HEV, PHEV, BEV, FCEV} is the fuel type of the car manufactured in year n.For each fuel type, we assumed that technological innovation improves LC-CO 2 emissions and ownership costs over the study period (see figure S1).The decision variables specify the number of ownership years of cars owned and is defined as: Here x n ∈ [0, X] is the number of years that a car manufactured in year n is owned; X is the maximum lifespan of a car.Using a case study, we identified the optimal replacement path (i.e.determining the fuel type required by the car, the fuel consumption rate, and when another fuel type should be used) for passenger car owners in Japan from 2005 to 2055.The optimisation problem for LC-CO 2 emissions was formulated as follows: Here, f (n) represents the minimum possible LC-CO 2 emissions accumulated from the start of year n to the end of year N, and x, t are the sets of variables obtained by solving the problem [10].The same procedure can be used to set up an optimisation problem for ownership costs.In the existing LCO model, when formulating an optimisation problem such as the one above, the analyst must assume a certain functional form for the objective function and variables.This can complicate the model design, for example, when considering irregular events (e.g. car inspection systems).We address this problem by solving the shortest path problem in graph theory with regard to the LCO model [14].Additionally, by leveraging the k-best algorithm [16,17], we can rapidly obtain multiple solutions to derive a robust replacement strategy.

Scenario analysis
Based on four scenarios with various parameters and uncertainties (table S3), namely the Current Trajectory Scenario (CTS), Business as Usual Scenario (BAU), Announced Pledges Scenario (APS), and Sustainable Development Scenario (SDS), we estimated the impact of exogenous factors, such as technological innovation, environmental improvements, and sustainability, on minimum emissions and costs.

Annual reduction of emissions
Based on the average annual emission reduction rate for 1990-2020 (Japan Automobile Manufacturers Association) [23], an annual emission improvement rate of 1.5% was assumed for the vehicle and battery production phases in the BAU scenario.The high level of the 3.0% annual emission improvement was based on the annual emission reduction rate for 2012-2020.Reducing energy use in the manufacturing process [24], building green supply chains [25], and product innovation [26] will help reduce the environmental impacts of the vehicle and battery production phases.

Annual reduction of vehicle and battery costs
Following the IEA [1] and ICCT [27], we assumed that vehicle and battery replacement costs fall for PHEVs, BEVs, and FCEVs as battery prices fall, owing to improvements in battery technology.However, GVs and HEVs are expected to maintain their sales prices until 2020.In addition, when the sales price of PHEVs falls to that of HEVs, the sales price does not fall any further.

Energy mix
The four scenarios for Japan's energy mix based on the IEA [28] are as follows: (1) the CTS, (2) the Stated Policies Scenario (STEPS), (3) the APS, and (4) the SDS.We assumed in this study that electricity is supplied directly during vehicle manufacturing and tank-to-wheel phases (e.g.power generation for EVs; equations (S4), (S5) and table S4).

Fuel efficiency improvement
In accordance with the Japan Automobile Manufacturers Association [29], improvements in the fuel efficiency of GVs, HEVs, and PHEVs will be maintained until 2030, 2035, and 2035, respectively.FCEVs are expected to maintain their current fuel economy until 2050 (table S5).In the BAU, the fuel efficiency improvement rate for each fuel type is maintained over the above period to achieve a fuel efficiency improvement of 32.4% in 2030 compared to that in 2016 [achieving the 2030 fuel efficiency standard (25.4 km l −1 ); Ministry of Land, Infrastructure, Transport, and Tourism] [30].For a higher level of fuel efficiency improvement, the target is revised upwards to a fuel efficiency improvement rate of 64.8% in 2030 compared to that in 2016 [achieving the 2030 fuel efficiency standard (31.6 km l −1 )].

Vehicle lifespan and battery replacement cycle
In this study, vehicle lifespan was defined as the period of ownership from new purchase to vehicle replacement [31].Following Nakamoto and Kagawa [32], the maximum lifetime of new passenger cars in the BAU scenario was set at 15 years.Lifetime extension is achieved through changes in consumer behaviour and the expansion of the repair market [11].
In the baseline scenario, we assumed that automobile batteries are replaced every eight years after new registration [3].In SDS, the battery replacement cycle is extended from 8 to 12 years by improving battery performance and durability.
The details and limitations of the methodology, equations, and data sources are provided in the Supplementary Information.

Lifecycle-CO 2 and ownership cost by fuel type
For an individual vehicle from 2020 to 2035, the lowest LC-CO 2 emissions and ownership costs are 22.3 tons for PHEVs and $30 861 for HEVs (figure 1).In terms of LC-CO 2 , EVs have low emissions during the tank-to-wheel phase but relatively high emissions during the manufacturing and battery replacement phases; therefore, improvements in emissions during these phases are necessary for effective reduction.Well-to-tank emissions are: GV: 6.5 tons (17.9% of the total), HEV: 3.3 tons (14.4%),PHEV: 7.1 tons (32.0%),BEV: 11.6 tons (46.7%), and FCEV: 17.8 tons (55.6%).There is room for significant emission reduction at this stage through green electricity generation and hydrogen production.
In terms of ownership costs, at current price levels, EVs are relatively expensive compared to GVs and HEVs, even with the use of subsidies.As vehicle and battery prices account for a relatively large portion of the ownership costs, this phase is critical for effective cost reduction.Fuel/electricity costs for driving are as follows: GV: $9958 (28.9% of total); HEV: $5137 (16.6%);PHEV: $4187 (12.6%);BEV: $4003 (9.7%); and FCEV: $9001 (14.8%), with EVs being relatively cheaper than GVs.

Optimal replacement path 2020-2055
In terms of LC-CO 2 from 2020 to 2055, the optimal emission reduction path was 45.1 tons with initial PHEV ownership.The replacement sequence is as follows: 1st car: drive PHEV for 14 years → 2nd car: drive PHEV for 15 years (until lifetime) → 3rd car: drive BEV for 7 years (figure 2(a)).For the other fueltype vehicles, the optimal (lowest LC-CO 2 ) path is to replace the PHEV as the second vehicle (in the 2030s) and the BEV as the third vehicle (after 2040), although the length of ownership of each vehicle slightly differs.This suggests that, given the current levels of technological innovation and emission improvements, a wide penetration of BEVs after 2040 is appropriate, and vehicle electrification should proceed gradually from GVs to PHEVs and then to BEVs.Except for GVs, HEVs, and FCEVs as the first cars, which were replaced relatively early (6 or 7 years after purchase), the other cars were owned for more than 14 years before being replaced, indicating that prolonged car use contributed to emission reduction.
Conversely, the replacement path with the lowest ownership cost spent $81 232 on initial HEV ownership (figure 2(b)).For each fuel-type vehicle except FCEVs, the optimal cost strategy is to replace it only with an HEV, followed by a BEV, as follows: 1st car: drive the initial car for 6 years; 2nd car: drive the HEV for 15 years (until lifetime); 3rd car: drive the BEV for 15 years (until lifetime).Similar to the emissions minimisation path, the cost minimisation path also favours keeping the vehicle for a relatively long time rather than replacing it in a short cycle.Upon analysis of the first through fifth optimal paths for LC-CO 2 and cost (table S6), the fundamental replacement strategy remained consistent and robust, although in some instances, vehicles were replaced with EVs a few years earlier or later, and in certain cases, replaced with PHEVs instead of HEVs.

Desirable replacement paths toward carbon neutrality
The effects of variations in each parameter on the selected vehicle's fuel type within the optimal replacement path and its ownership duration, along with their consequential impacts on emissions and costs, diverge (figure 3).For LC-CO 2 , the change in emissions associated with changes in travel distance is about ±30%, which is relatively larger than the impact of changes in other parameters.The impact of changes in price and lifespan on cost is ±20%, which is relatively larger than the impact of changes in other parameters.
Vehicle usage conditions by car owners are expected to include a range of favourable to unfavourable cases.In addition, the parameters in the sensitivity analysis are projected to trend  favourably above the baseline level in the future [1].Consequently, as changes in vehicle owner usage conditions align with desirable shifts in the parameters, higher levels of electrification will emerge as the optimal strategy across all scenarios in this study (table S7).
Enhanced energy efficiency and decarbonisation in the power sector will accelerate the timeframe for BEV deployment to reduce emissions (figures 4(a)-(d)).For example, comparing BAU and APS for initial PHEV ownership, BAU switched to BEV after 30 years, whereas APS switched to BEV earlier (in 16 years) and reduced emissions by 7.2 tons (−16.0%relative to BAU; figure 4(b)).In addition, under SDS, the number of replacements over the analysis period was reduced from two to one, and LC-CO 2 emissions were reduced by 12.8 tons (−28.3%)compared to that for BAU.To reduce emissions further, improvements in energy efficiency and green power generation must be accompanied by long-term product use.However, under the CTS, the optimal path is to replace the second and third cars with HEVs, regardless of the fuel type of the first car.The results suggest that continued improvements in emissions and decarbonisation of the power sector are essential for vehicle electrification to reduce emissions, and if these In terms of the cost of ownership, the optimal strategy under BAU is to switch to HEVs, although a path exists where switching to BEVs is optimal as the cost of ownership of EVs declines (figures 4(e)-(h)).For example, under APS, an owner who owns a HEV as their first car would be best served by replacing it with BEVs as their second and third cars, saving $14 675 (−18.1%)compared to the cost under BAU (figure 4(g)).In addition, under SDS, the reductions for each fuel type compared to those for BAU are: GV: $18 811 (−22.9%),HEV: $22 065 (−27.2%),PHEV: $21 550 (−26.4%), and BEV: $20 405 (−24.4%).Under the CTS, as under the BAU, the optimal path is to replace the second and third vehicles with HEVs, regardless of the fuel type of the first vehicle.
Compared to baselines that assume longer vehicle use, as recommended in existing studies [32][33][34], the optimal replacement strategy based on this study resulted in average emission and cost savings of 9.0 and 3.6%, respectively (table S8).For certain fuel types, transitioning to HEVs or EVs relatively early in the vehicle's lifespan, rather than using the vehicle for a long period, may contribute to emission and cost savings.

Prolong use of older car or replace?
To optimise emissions and costs, vehicle owners would be better off replacing their vehicles with energy-efficient HEVs and PHEVs than continuing to drive older, less-energy-efficient GVs for longer periods (figure 5).In the emission optimisation case, under BAU (figure 5(a)), GVs manufactured and purchased in 2005, 2010, 2015, and 2020 (05, 10, 15, and 20 opt) were maintained for one, one, one, and six years, respectively, and then replaced with HEVs or PHEVs.For GVs manufactured and purchased in 2005, 2010, and 2015, if the replacement path is optimised after 2020 with no replacement by 2020 (05 f20, 10 f20, and 15 f20), the GVs will be replaced by PHEVs in 2020.As vehicle owners replace their vehicles, emissions can be optimised by switching to HEVs by 2020, PHEVs from 2020 to 2040, and BEVs thereafter.Similarly, in the SDS, it is preferable to replace the first GV with an HEV or PHEV as early as possible and maintain the vehicle for a long time.
Similarly, in the cost optimisation case, after owning a GV for a relatively short period, it is desirable to drive an HEV as the second and third vehicle and maintain the vehicle for a long time.Under the BAU (figure 5  For example, concerning GVs manufactured and purchased in 2005, the optimal replacement strategy indicated in this study (05 opt) is more effective in terms of LC-CO 2 and cost than the long-term use of GVs (05 f20) under BAU, resulting in savings of 10.5 tons (12.3%) and $3588 (2.9%) in costs (table S9).
A key finding of this study is that for many regions, replacing older, less energy-efficient GVs on the road with energy-efficient HEVs and EVs, rather than prolonging their use as suggested by existing studies [32][33][34], will substantially contribute to emission and cost reductions.

Policy implications
In the case of emission optimisation, a broad penetration of BEVs after 2040 is appropriate under the BAU scenario, and vehicle electrification should proceed gradually from GVs to PHEVs and then to BEVs (figure 2).However, in the APS, BEVs should be introduced in the 2030s to reduce emissions, a decade earlier than in the BAU scenario (figure 4).In the BEV transition policy, policymakers in regions that are highly dependent on fossil fuels should encourage a gradual transition to fuel types such as HEVs and PHEVs, which are superior to BEVs in terms of emissions and cost, until they achieve a high level of LC-CO 2 emission reduction in cars and create an environment in which BEVs should be introduced.
As driving emissions account for a large share of vehicle LC-CO 2 , approaches tailored to car owners are also important for efficient emission reduction (figures 1 and 3).For example, reducing the share of private cars in transportation (e.g. using public transportation and promoting car-sharing services) can reduce private car journeys [35,36].Furthermore, in addition to traditional environmental measures related to automobiles, such as setting fuel efficiency standards, green power generation will play a crucial role in reducing emissions associated with the widespread use of EVs.
Even when considering equal emission improvements and cost reduction, it is preferable in terms of both emissions and costs to extend the ownership of a vehicle rather than opting for frequent replacement.In this regard, our study aligns with existing literature on this subject [32][33][34]37].However, there is potential for extended car ownership in some countries and regions; for example, the average ownership period of new cars (i.e. the average number of years between new registration and deregistration) in Japan in 2020 was 13.8 years [38], which is lower than that in other countries (e.g.Australia, Brazil, Finland, and the U.S.) [39].Improving the durability and reparability of products [11] and expanding the second-hand market [40] are critical to facilitating the long-term use of products.
A key insight from this study underscores the selective and extended use of energy-efficient vehicles rather than the uniform utilisation of all vehicles.Owners of less energy-efficient vehicles should consider replacing them with more efficient alternatives rather than maintaining their current usage.Presently, the Japanese government incentivises older vehicle owners (exceeding 13 years from initial registration) to replace their cars by applying a 15% annual vehicle tax surcharge [41].Certain studies have proposed that more effective emission reduction could be achieved by restricting age and fuel efficiency criteria for vehicle subsidies aimed at promoting energyefficient options [33,42].Thus, the strategic design of a system that prioritises the replacement of less energy-efficient vehicles becomes crucial for successful vehicle electrification [43].
Emissions and cost optimisation are incompatible with vehicle replacement paths.Therefore, the appropriate vehicle replacement cycle and fuel-type depend on what is optimised.A roadmap to effectively reduce vehicle emissions (and costs) must bridge the gap between the perspectives of policymakers and environmentally conscious car owners who prioritise emission reductions and those who prioritise cost reduction.For example, the implementation of an eco-label (revealing the cost-effectiveness of switching from a conventional car to an EV) can show consumers the cost of the vehicle based on a long-term perspective (including external costs due to environmental problems) rather than the cost in terms of the current price.

Limitations and future directions
This study has a few limitations, some of which can direct future research.As this study assumed that altering the energy mix has no influence on electricity prices, there was no observed impact on ownership costs.However, this issue presents a challenge that forthcoming research endeavours must tackle.Furthermore, cost is a factor related to consumer utility, and we do not consider other traditional attributes and behavioural components.Unless there are no deviations between the ideal path and the behaviour of car owners, the real-world implementation will differ from the optimal roadmaps.The analytical models to address these challenges would increase complexity in the configuration of the optimisation problem.In this context, the shortest path approach [14] is more useful.
In this study, we assumed that battery costs and car sales prices will follow an exogenously determined path, which ignores important feedback on market formation.As more EVs are sold, prices are expected to decline more rapidly due to economies of scale and learning curves [1].Moreover, increased sales will expand the range of makes and models within each vehicle type.Similar dynamics are applicable to complementary assets, such as EV home chargers and repair shops.Conversely, factors such as consumer perceptions of EVs, feedback affecting market formation, falling prices of substitutes, and resource availability may hinder the pace of EV penetration.These dynamics and constraints may reshape the roadmap presented in this study.
One promising avenue for future exploration involves expanding the model beyond individual car owner replacement paths to optimise the transition journey towards vehicle electrification on a country or regional scale.Notably, integrating resource availability and national electrification objectives as constraints could yield more pragmatic transition strategies.Alternatively, employing the k-best algorithm [14] to derive multiple suboptimal paths could offer policymakers and car owners a more practical selection, such as a path expected to yield substantial emission reduction (cost savings) with less intensive efforts than the optimal replacement trajectory.Although the k-best shortest paths approach provides several useful optimal strategies, it is not possible to incorporate hard constraints such as budget constraints.In this case, a dynamic programming approach may also be useful.

Figure 1 .
Figure 1.Lifecycle-CO2 and ownership cost of cars.LC-CO2 emissions (a) and ownership costs (b) by fuel type.Assumptions regarding vehicles include an annual mileage of 10 000 km, vehicle lifetime of 15 years, and one battery replacement in HEVs, PHEVs, BEVs, and FCEVs during their lifetimes.The results are based on the energy mix from 2020 to 2035.Crosses indicate net ownership costs.

Figure 2 .
Figure 2. Optimal replacement path by fuel type.Replacement paths that optimise LC-CO2 emissions (a) and ownership costs (b) are shown.

Figure 3 .
Figure 3. Sensitivity analysis.LC-CO2 changes (a) and ownership cost changes (b) with varied parameters.+ indicates fuel-type changes, and * denotes shifts in ownership periods.Parameters in favourable and unfavourable cases use Sustainable Development Scenario and Current Trajectory Scenario values, respectively.Note: Favourable case travel distance is 5000 km yr −1 ; unfavourable case has a 15 000 km yr −1 travel distance, 10 year vehicle lifespan, and 4 year battery replacement cycle.
(d)), the GVs manufactured and purchased in 2005, 2010, 2015, and 2020 (05, 10, 15, and 20 opt) were maintained for 6, 1, 12, and 6 years, respectively, and then replaced with HEVs.Although the existence of a pathway to replace GVs with energy-efficient GVs manufactured in 2015 or 2020 or to keep GVs relatively long is a feature of the pathway in terms of cost optimisation, older GVs were replaced with energyefficient vehicles earlier.

Table 1 .
Specifications and cost of cars by fuel type.
a Tax benefits apply to new car purchases and initial inspections, excluding GVs intended for fuel efficiency.bAfter13 years, GVs face a 15% surcharge.cTheGreening Special Exception offers a 75% tax reduction two years after registration.d Ongoing inspections are required for subsequent assessments.