Green Last-Mile Delivery: Adapting Beverage Distribution to Low Emission Urban Areas

Abstract

Electrifying urban last-mile logistics is an important step towards reducing carbon emissions which requires replacing conventional vehicles with low-carbon alternatives that offer comparable operational and cost characteristics. This study presents a methodology for evaluating the feasibility of electrifying an urban delivery fleet, using data from a major beverage company in Seville as a case study. Applying a fleet and route optimization algorithm for various vehicle combinations, we demonstrate that emerging electric vehicle options, combined with a redesigned fleet mix and an optimized routing, can already enable cost-efficient electrification of distribution activities in the city centre. Furthermore, our analysis suggests that full electrification of the company’s local distribution network may be possible by 2030, depending on the availability of larger electric trucks. Our results show that currently available electric vehicles can fully substitute conventional options in the case study context, with higher capital costs offset by lower energy costs in most cases. The electrification of urban logistics can yield significant environmental benefits, particularly if powered by a clean energy mix.

1. Introduction

The city logistics sector plays a vital role in supporting urban economic activities, but also poses significant environmental and social challenges, including congestion, air pollution, noise, and road accidents. In Europe, city logistics is responsible for 25% of urban transport-related CO₂ emissions and 30–50% of other transport-related pollutants [1]. The growing urban population, combined with trends such as the rise in e-commerce and home delivery, is expected to increase demand for goods and services, leading to a surge in urban freight demand.

In response to these challenges, the European Commission has adopted several initiatives aimed at reducing emissions from the transport sector. The Sustainable and Smart Mobility Strategy [1] and the Action Plan for sustainable mobility outline a comprehensive approach to achieving a “green and digital transformation” of the EU transport sector, in line with the goals of the European Green Deal. The main goal is to provide a strategy to accomplish the “green and digital transformations” of the EU transport sector and meet the goals as set out by the European Green Deal.

The New Urban Mobility Framework [2] further emphasizes the importance of “zero-emission city freight logistics and last-mile deliveries”, highlighting the need for low-carbon vehicles, novel distribution models, and sustainable urban logistic plans. This new proposal includes actions ranging from the inclusion of low-carbon vehicles to the adoption of novel distribution models, the promotion of sustainable urban logistics plans (SULPs) and the cooperation between public and private stakeholders, including voluntary data sharing [3].

Improving urban freight transport planning is a complex task, particularly for cities seeking to reduce their environmental footprint. Captive fleets, such as those used by beverage and food delivery companies, offer a promising opportunity for early adoption of alternative fuel or vehicle types, reducing emissions, noise, and congestion externalities.

This case study explores the potential for a captive fleet to contribute to achieving climate neutrality in urban delivery systems. Using realistic data from delivery operations by a major beverage producer in Seville, we simulate possible scenarios for reducing emissions and improving operational efficiency. The findings of this study provide valuable insights into the opportunities and challenges associated with adapting urban delivery systems to meet the climate goals of both cities and companies.

Several studies evaluated the potential introduction of low-carbon vehicles in urban freight transport, covering a wide range of issues from operations (routing problems and technical feasibility) to impacts (economic and environmental aspects, policy options) [4]. Perboli and Rosano [5] simulated emissions and costs of implementing specific scenarios of parcel deliveries, using combinations of low-carbon vehicles and diesel vans in Turin, Italy. Their model assumes the presence of a local/secondary (urban) depot for commercial cargo bicycles, however only CO₂ emission externalities are considered. The authors use synthetic distributions of delivery demand and routing, informed by specific parameters from the city case study and data from the URBeLOG project [6,7]. As regards the behaviour of urban freight operators, Stockhammer et al. [8] provide an overview of factors that influence their choice when selecting vehicle modes.

Settey et al. [9] and Giordano et al. [10] explored the effects of topography, driving cycles, and load weight/volume on the operational feasibility of electric vehicles and bicycles in urban environments, respectively. However, their findings are focused on specific vehicle technologies, rather than on the operational scenarios that companies could implement to include low-carbon vehicles into their existing fleets. Melo and Baptista [11] developed a traffic simulation model to estimate operational and environmental benefits of electric cargo bi/tricycles, comparing their energy savings to diesel vans in Porto, Portugal. Other studies focus on battery electric vehicle (BEV) energy consumption using GPS data from delivery operations to serve as inputs in their route optimization or vehicle assessment models [12].

While the existing literature on urban freight transport and last-mile delivery has highlighted the importance of reducing emissions and environmental impacts, there are still several gaps and limitations in the research. Many studies have focused on the technical feasibility of alternative fuel vehicles, such as electric or hybrid vehicles, without adequately considering the operational and economic implications of their adoption [13,14,15,16,17,18,19]. For instance, the study by Galati et al. [13] found that electric vehicles can be a cost-effective option for short food supply chains, but their analysis was limited to a specific context and did not account for the variability in demand and delivery routes.

Furthermore, the literature has often relied on simplified assumptions about the behaviour of freight operators and the characteristics of urban freight demand. For example, the study by Perboli and Rosano [5] simulated the emissions and costs of parcel deliveries using a combination of low-carbon vehicles and diesel vans, but their model assumed a fixed distribution of delivery demand and did not account for the potential impacts of traffic congestion or road network characteristics.

In addition, the existing research tends to focus on specific aspects of urban freight transport, such as vehicle technology or routing optimization, without considering the broader systemic implications of different logistics strategies [20,21]. This limited the scope of the analysis to a specific situation and did not consider the potential scalability or transferability of their findings to other contexts.

Several studies have assessed road transport external costs, such as congestion, noise, road accidents, and road damage. Most focus on conventional passenger cars at country-level, and their cost estimates are not directly applicable to cities and commercial vehicles. Reference [22] examined the environmental impact of urban road freight in Paris, using models to estimate traffic flows and assess pollutant emissions, finding significant environmental costs attributed to freight traffic in the region. However, external costs are limited to airborne emissions, and data availability limitations for freight demand modelling reduce the accuracy and the applicability of the findings. Differentiations of external costs per road vehicle types are also present in [23,24,25]. However, the level of geographic aggregation and the limited coverage of commercial vehicles limit the application of their estimates to specific cities.

In spite of the great variety of scientific literature dealing with last-mile deliveries, few studies directly address externalities and costs faced by delivery fleet operators, and even fewer applied their methodologies to model realistic low-carbon vehicle fleet scenarios in urban environments. To address these limitations, our study aims to provide a more comprehensive and integrated analysis of the potential for electrifying urban delivery fleets, taking into account the complex interactions between vehicle technology, logistics operations, and urban transport systems. By combining a critical review of the existing literature with empirical analysis and scenario-based modelling, we seek to contribute to a deeper understanding of the challenges and opportunities associated with reducing emissions and environmental impacts in urban freight transport.

This case study aims to contribute to the field by applying a scalable approach, based on data from real operations. It does so by comparing the external and private costs caused by the implementation of different scenarios of fleet composition (i.e., including BEV and diesel vans, electric cargo tricycle, or cargo bicycles) in the case of a major beverage delivery company in the city of Seville, Spain.

2. Materials and Methods

Last-mile distribution is important in terms of urban transport activity and its impacts, but an optimal balance between economic efficiency and environmental performance is often difficult to achieve. Given the ambitious strategies of the EU [2], a growing number of cities are exploring strategies to reduce GHG emissions from urban logistics [26]. In the private sector, several operators are aiming at carbon neutrality, from production to distribution, as well as at testing innovative pilot applications that facilitate the introduction of electric micro-vehicles. In the work presented here, we analyze a case of adapting last-mile distribution to the requirements of a new Low Emissions Zone (LEZ) in the city of Seville. The case study was carried out in collaboration with a major beverage company operating in the city, through the exchange of data and expertise concerning the operational requirements by the company. The analysis compares alternative options for the adaptation of the company’s urban distribution system to the new urban policy requirements. The approach uses suitable economic and environmental indicators that address both the operator’s and the city’s priorities.

The data provided by the company cover the distribution demand and routes in Seville and its suburbs, covering the beverage distribution operations during a typical day (see

Table 1. Data provided by the beverage company operating in Seville for a typical business day.

Points of Delivery (POD) Address	Routes (n = 8)
Starting point (SP) address	Route number
Time slot of delivery (start/end)	Starting point (warehouse address/coordinates)
Estimated delivery duration	Start/end time
Delivered quantity (‘equivalent packages’ or kg)	Total distance driven
Geo coordinates (POD, SP)	Total quantity delivered (‘equivalent packages’ or kg)
Route number	Vehicle type
Sequence number in route	Maximum vehicle capacity (‘equivalent packages’ or kg)

). The data include 141 delivery points, and (as multiple delivery orders are in the same place) 106 delivery stops. Eight routes were included, with one of them having its delivery points in the future Seville Low Emissions Zone (LEZ). Figure 1a illustrates the points of delivery and the current warehouse/distribution centre and Figure 1b zooms into the LEZ.

Based on data availability and interactions with distributors, we applied tailored fleet management strategies and routing models to test options for the gradual decarbonisation of the distribution routes, and to assess specific scenarios in terms of their economic and environmental impacts. The application was developed using an implementation of the OpenRouteService [27] in R and was applied for the combined optimization of fleet mix and route choice. We tested the scenarios outlined in

Table 2. Scenario descriptions.

Scenario	Scenario Group	Scenario Description
C	Baseline	Full conventional fleet
CEZ	Vehicle Fleet	Conventional routes outside Low Emission Zone (LEZ), BEVs in LEZ
CTEV	Vehicle Fleet	Conventional trucks, BEV vans
E	Vehicle Fleet	All BEVs [vans and trucks]
DEZ1	City Warehouse + Partial Vehicle Fleet	Conventional trucks, BEV vans and cargo tricycles in LEZ/City Centre
DEZ2	City Warehouse + Partial Vehicle Fleet	Conventional trucks, BEV vans and cargo tricycles in LEZ/City Centre, bicycles in LEZ
DEV1	City Warehouse + Partial Vehicle Fleet	Conventional trucks, BEV vans, plus BEV vans and cargo tricycles in LEZ/City Centre
DEV2	City Warehouse + Partial Vehicle Fleet	Conventional trucks, BEV vans, plus BEV vans and cargo tricycles in LEZ/City Centre, bicycles inside LEZ
DET1	City Warehouse + Vehicle Fleet	ALL BEVs and cargo tricycles in LEZ/City Centre
DET2	City Warehouse + Vehicle Fleet	ALL BEVs and cargo tricycles in LEZ/City Centre, bicycles in LEZ

, including several variables such as time horizons (2025 or 2030), combinations of vehicle technologies (diesel/electric trucks and vans, electric tri/bicycles), geographical coverage limitations (full distribution, or electric inside LEZ/City Centre and conventional outside), and existence of distribution centres (a currently used distribution centre, or an additional warehouse located in the LEZ perimeter).

In each scenario, after accounting for the vehicles’ size, their carrying capacity and speed, we estimated the required number of vehicles by powertrain and size that can cover demand at optimal efficiency. In assessing technological options, we made forecasts based on desk research and existing literature to identify potential options for electric trucks, vans, and cargo bicycles by 2030. This process included the analysis of factors such as the vehicles’ purchase and operation costs, energy consumption, carrying capacity, speed, and range. Additionally, the analysis incorporated cargo capacity and energy consumption data for both conventional and emerging vehicle technology options available in the market or close to their commercialization.

The analysis and comparison of the specific scenarios were conducted using indicators that enabled the assessment of operational, economic, and environmental impacts. For each scenario, the following indicators were estimated:

• Total driving distance (vehicle/km) per technology (conventional/electric)
• Number of vehicles required (by vehicle technology and size)
• Fuel/electricity energy consumption (conventional/electric)
• Total cost of ownership (by vehicle technology and size) not including the cost of the new warehouse
• Total CO₂ emissions (by vehicle technology and size)
• Total external costs (by vehicle technology and size)
• Total manpower required (by vehicle technology and size)

The costs used for the calculations were based on assumptions concerning vehicle purchase, maintenance and depreciation costs, and on personnel costs. Electric vans capital costs were assumed to be higher than diesel vans (40% higher in 2025 and 20% higher in 2030 due to battery price evolution). This is a more conservative approach than in a recent study by [28], where the purchase price parity between diesel and electric vans is expected earlier. Price differences between electric and diesel trucks are based on [29]. Electric truck purchase costs were 200% higher in 2025 and 70% higher in 2030 compared to diesel trucks. The intermediate distribution point (city warehouse) is a potential investment made by the city of Seville and would not have a direct repercussion on the costs of the operators.

The required energy consumption was estimated based on the vehicle technology, size, and age options, assuming flat topography and using COPERT v.5.5 software for diesel vehicles, and results from the Handbook on External Costs of Transport [30] for low-carbon vehicle options. GHG emissions per kilowatt/hour were estimated based on the electricity generation mix at country level from the European Environment Agency [31]. In addition to the estimates on fuel consumption and emissions, the approach includes the assessment of fixed and variable private costs, including the potential impact on labour demand. Annual mileage projections for the fuel/energy costs are based on the typical day of operations provided by the beverage company and 300 business days per year, while for capital costs we assume a vehicle life of twelve years. Finally, the external costs of transport are calculated for the different scenarios. Besides GHG emissions, these include noise, air pollution, road damage, congestion and road accidents, based on [25].

3. Results

The first question concerning the potential shift to alternative technologies addresses the changes in the fleet mix necessary in each scenario. The main factors that affect the number of vehicles required are the capacity and range for each alternative.

Table 3. Distance per day (km) and type of vehicles employed in each scenario. Vehicles description includes payload.

Vehicle Type and Payload	C	CEZ	CTEV	E	DEZ1	DEZ2	DEV1	DEV2	DET1	DET2
Conventional truck 6.5 t	82	82	82	-	82	82	82	82	-	-
Conventional van 1.4 t	260	152	-	-	152	152	-	-	-	-
Electric truck 6.5 t	-	-	-	82	-	-	-	-	82	82
Electric van 1.4 t	-	108	260	260	71	71	222	222	222	222
Electric cargo tricycle 0.5 t	-	-	-	-	21	19	21	19	21	19
Electric cargo bicycle 0.1 t	-	-	-	-	-	12	-	12	-	12
Total	342	342	342	342	325	335	325	335	325	335

presents a summary of the distances by each vehicle type, and scenario, and it reports the maximum cargo weight capacity of each vehicle. Most of delivery points in the City Centre included in Figure 1b are well within the delivery range of the tricycle, including those outside the Low Emissions Zone. When analyzing the delivery range of the e-bicycle, we took into account delivery points located within the Low Emissions Zone next to the new warehouse.

Table 4. Number of vehicles per scenario.

Vehicle Type and Payload	C	CEZ	CTEV	E	DEZ1	DEZ2	DEV1	DEV2	DET1	DET2
Conventional truck 6.5 t	2	2	2	-	2	2	2	2	-	-
Conventional van 1.4 t	6	5	-	-	5	5	-	-	-	-
Electric truck 6.5 t	-	-	-	2	-	-	-	-	2	2
Electric van 1.4 t	-	1	6	6	1	1	6	6	6	6
Electric cargo tricycle 0.5 t	-	-	-	-	1	1	1	1	1	1
Electric cargo bicycle 0.1 t	-	-	-	-	-	1	-	1	-	1

indicates the number of vehicles needed (by technology) in each scenario. In the specific case study assessed in this paper, which included data from a single “typical” day of operations of the beverage company in Seville, the deliveries performed by the tri/bicycle do not reduce the number of vans required. It is expected that a more comprehensive analysis of routes over a more extended period will help to better characterize the delivery operation, which could result in a reduction in number of vans in the fleet in favour of tri/bicycles. The required Full Time Equivalent (FTE) personnel remain the same in all scenarios, since just one full-time van driver is converted into half-time, and only an additional half-time driver for tri/bicycle is required.

A summary of annual costs is presented in

Table 5. Fleet annualized costs in 2025 and 2030.

Cost (‘000s EUR/Year)	C	CEZ	CTEV	E	DEZ1	DEZ2	DEV1	DEV2	DET1	DET2
Vehicles fixed costs 2025	21.9	22.9	27.8	42.8	23.5	23.9	28.4	28.7	43.4	43.7
Vehicles fixed costs 2030	21.9	22.4	24.6	29.9	22.8	23.2	25.1	25.4	30.3	30.7
Fuel/energy costs 2025, 2030	32.3	25.9	16.8	12.1	25.1	25.1	16.1	16.1	11.3	11.3
Personnel cost 2025, 2030	223.4	223.4	223.4	223.4	223.4	223.4	223.4	223.4	223.4	223.4

for 2025 and 2030. Fuel, electricity, and personnel costs are kept constant between 2025 and 2030 to capture the impact of vehicle capital costs. While we find that more than 80% of the costs are related to personnel (truck and van drivers’ wages and insurance costs are assumed to be the same of tri/bicycles riders). Moreover, the higher capital costs of electric vehicles, compared to their diesel counterparts, are offset by the reduced electricity expenses in comparison to fuel costs. These findings are in line with [13,16,17] and are justified by the projected high annual mileage of the vehicles in the fleet to meet the beverage distribution demand. Similarly to these studies, our results also reveal that fleets with battery electric vehicles are cost-competitive with their diesel fleet counterparts. Differently from the above-mentioned literature, our results are applied to fleets scenarios, simulated over real delivery routes in urban settings. The findings aim to contribute to the discussion on how to include low-carbon emission vehicles in green fleets and on their expected economic and environmental impacts.

Table 6. Scenarios’ CO₂ emissions, cost per day, and comparison to Scenario “C” in 2025 and 2030.

Year 2025	C	CEZ	CTEV	E	DEZ1	DEZ2	DEV1	DEV2	DET1	DET2
Emissions (tCO2/year)	63	47	24	7	46	46	24	24	7	7
Cost (‘000s EUR/year)	278	272	268	278	272	272	268	268	278	279
Emissions change	-	−25%	−61%	−89%	−26%	−26%	−62%	−62%	−89%	−89%
Cost changes	-	−2%	−3%	0%	−2%	−2%	−4%	−3%	0%	0%
Year 2030
Emissions (tCO2/year)	63	46	23	5	46	46	23	23	5	5
Cost (‘000s EUR/year)	278	272	265	265	271	272	265	265	265	265
Emissions change	-	−26%	−63%	−92%	−27%	−27%	−63%	−63%	−93%	−93%
Cost changes	-	−2%	−5%	−4%	−2%	−2%	−5%	−5%	−5%	−4%

presents the annualized emissions and costs for 2025 and 2030. The most significant CO₂ emissions reductions compared to Baseline scenario ‘C’ appear when the complete van fleet is electric, with reductions greater than 60% and reaching 90% when the truck fleet is converted to electric. On the other hand, the partial electrification of vans (i.e., only those crossing the LEZ/City Centre) shows a limited CO₂ emission decrease of 25%. Costs of the different scenarios compared to Baseline scenario ‘C’ are in the range of [−4%, 0%] for 2025 and [−5%, −2%] for 2030. In the latter, lower capital costs are assumed for electric vehicles compared to 2025.

In order to test the robustness of the results, we carried out a sensitivity analysis using modified distributions of the delivery points across the urban area. We followed the approach described in [32] and geo-localized 60 bars and restaurants within the LEZ that could potentially be a client of the beverage distribution system. We performed 500 simulations, each using a random subset of 20 delivery points. Similarly to the approach in [32], the selection of 20 delivery points in the study is considered as a realistic pattern for beverage distribution, an assumption confirmed by interviews with operators. In each simulation, the stochastic selection of delivery points resulted in varying average distances to the main distribution centre and to the intermediate distribution centre (city warehouse). As a consequence, the optimal combined fleet mix and routing solution identified by the OpenRouteService algorithm produced different estimates for total costs and emissions in each scenario.

Figure 2 is an illustration of the variance of the estimates, using the example of scenario DET2. In 369 of the 500 simulations (73.8%), the total costs for scenario DET2 in 2030 are estimated to be lower than those for the baseline scenario C, i.e., a shift to a combination of BEV’s, cargo tricycles and bicycles in the LEZ with an intermediate distribution centre would result in an economic benefit for the operator. The median corresponds to an annual benefit of EUR 17,600, with a mean of EUR 13,800, maximum of EUR 48,100 and minimum of EUR −32,500 (loss). The x-axis in Figure 2 represents the ratio of the total route distance within the zone served through the intermediate distribution centre, compared to the total route distance if all deliveries are made directly from the main distribution centre. A higher ratio corresponds to the intermediate distribution centre being closer to the main distribution centre and –as a consequence- at a longer average distance from the delivery points. The y-axis represents the ratio between the time required to cover each route by van and that by electric bi/tri-cycle. This ratio captures the difference in speed between the two modalities and can be a proxy for the difference in driver productivity. The size of the bubbles in the graph represents the total benefit or loss for the operator in scenario DET2, compared to the baseline C. It is obvious from the graph that there is a sweet spot where scenario DET2 is consistently more beneficial than the baseline scenario. Most simulations that estimate a benefit are concentrated towards cases where the delivery destinations are located at shorter distances from the city warehouse and can be accessed through the city’s cycling network. In such cases, cargo bi/tri-cycles may be more efficient than delivery vans, despite their lower speed and capacity. The cases where the combination of a city warehouse with bi/tri-cycles may operate at a loss tend to concentrate in two regions of Figure 2: towards the right side of the figure, in cases where the delivery points are spread widely across the city, direct distribution from the central distribution point may be more efficient than using the city warehouse as an intermediate distribution centre; and towards the bottom left of the figure, in cases where the delivery points are located in less dense areas, the operational efficiency of bi/tri-cycles is lower than that of delivery vans.

Figure 3 summarizes the external cost estimates in the 2025 scenario. The estimates are a combination of fleet composition described in Table 3 and Table 4 with vehicle externalities’ factors on a per vehicle/kilometre basis. The baseline diesel fleet, corresponding to scenario ‘C’, is the baseline where all deliveries are performed by vehicles with Euro II to Euro V emission standards (i.e., we take the mean emission estimates between these two standards). In this scenario, we estimate that the annual external costs of operating two trucks and six conventional diesel vans, to meet the beverage demand of business in the metropolitan area of Seville, could vary between EUR 75,000 and 115,000. The main factors contributing to these costs are congestion, air pollution, noise and GHG emissions. The analysis finds that the main external cost reductions, derived from introducing BEV and tri/bicycles and compared to the baseline scenario, are in terms of GHG emissions, air pollution and noise. These environmental benefits are more pronounced in scenarios “E”, “DET1” and “DET2”, where internal combustion engine vehicles are entirely excluded. These are followed by scenarios “CTEV”, “DEV1” and “DEV2”, which permit the incorporation of conventional trucks in the fleet mix only for operating delivery routes in the city outskirts.

Since electric and conventional vans and trucks have practically the same size, we observe a benefit in terms of the costs of congestion only when the fleet includes lighter vehicles. However, these positive effects are constrained by the limited mileage tri/bicycles could cover in the scenarios, given the combination of size and weight of the beverage goods and the limited load capacity of these vehicles. At an aggregate level, the case study estimates that benefits, in terms of reduced externalities, can amount to up to EUR 40,000–70,000 annually. The outcomes can also be categorized into three tiers: (i) a low reduction in external costs (from −10% to −15%) when only vans entering the LEZ/City Centre are electrified; (ii) a medium reduction (from −25% to −30%) when all vans are converted to electric; and (iii) a high reduction (from −37% to 42%) when the entire fleet is transformed to either electric vehicles or a combination of BEVs and tri/bicycles.

4. Discussion

The results of this case study contribute to the discussion on the potential of emerging technologies for green distribution and their expected impacts in economic and environmental terms. Our findings on the economic viability of electrifying urban delivery fleets are in line with those of Gil Ribeiro and Silveira [16], who also conclude that fleets with battery electric vehicles are cost-competitive with their diesel fleet counterparts. However, our results are applied to various fleet combination scenarios simulated over real delivery routes in urban settings, extending the scope of the analysis to a combined fleet mix and logistics optimization problem. In contrast to the study by Galati et al. [13], which found that electric vehicles were not cost-effective for short food supply chains, our results suggest that the higher capital costs of electric alternatives can already be compensated by the lower energy costs in most cases.

The environmental benefits of electrifying urban delivery fleets are also consistent with the findings of other studies [33]. For example, the study by Iwan [34] found that electric mobility can reduce emissions in urban freight and logistics, while the study by Napoli et al. [35] found that freight distribution with electric vehicles can reduce emissions and operating costs. Our results also corroborate the findings of Ehrler et al. [36], who found that electric vehicles can be a viable option for last-mile logistics of grocery e-commerce. However, our study goes further by considering the potential impacts of different logistics strategies, such as the introduction of intermediate distribution centres, on the economic and environmental performance of urban delivery fleets.

The boundary conditions for the use and efficiency of an intermediate distribution centre in this case study are consistent with the theoretical expectations outlined in Perboli and Rosano [5]. Our findings on the potential of intermediate distribution centres to facilitate the use of smaller and cleaner vehicles for last-mile distribution are also supported by case studies on urban logistics, such as the study by Bruni et al. [37], Savall-Maño and Ribas [38] and Katsela et al. [39]. However, our results highlight the importance of considering the average weight and volume of individual shipments, as well as urban density and average distribution distance, when evaluating the suitability of cargo bi/tri-cycles. This is in line with the findings of Schliwa et al. [40], who found that cargo cycles can be an efficient option for last-mile delivery in urban areas but require careful consideration of the operational and logistical constraints.

The end-use of electric mobility is a critical component of the overall electromobility ecosystem. As the demand for electric vehicles grows, the development of charging infrastructure and the production of electric vehicles will need to keep pace. However, the end-use of electric mobility is also where the benefits of electrification are most directly realized, in terms of reduced emissions and improved air quality. By focusing on the end-use of electric mobility, our study is able to provide insights into the ways in which electric vehicles can be used to improve the efficiency and reduce the environmental impacts of urban freight transport. We find that electrifying urban delivery fleets can have significant economic and environmental benefits, including reduced fuel costs, lower emissions, and improved air quality. These benefits are consistent with the findings of other studies, which have highlighted the potential for electric vehicles to reduce emissions and improve air quality in urban areas [16,34,35].

Overall, our study contributes to the existing literature by providing a more comprehensive and integrated analysis of the potential for electrifying urban delivery fleets, taking into account the complex interactions between vehicle technology, logistics operations, and urban transport systems. By comparing our results with those obtained by other scientists, we can see that our findings are consistent with the broader trends and patterns observed in the field but also provide new insights and perspectives on the challenges and opportunities associated with reducing emissions and environmental impacts in urban freight transport. Furthermore, our study’s findings are supported by [41], which applied a quantitative scenario-based model for assessing the impacts of cargo bike transhipment points in urban districts. In addition, the study’s findings are in line with the results of [42], which investigated the vehicle routing problem with delivery options and found that cargo bikes can be a viable option for reducing emissions in urban areas.

A particular aspect that needs to be taken into account for the interpretation of the results is the fact that the intermediate distribution centre—a main element for the design of the green last-mile delivery scenarios—is assumed to be provided by the city of Seville free of cost to the operator. The measure was adopted by the city of Seville as part of its strategy to achieve carbon neutrality in urban logistics [43]. This is obviously a site-specific situation which cannot be generalized for any similar application elsewhere. In the case where the establishment of an intermediate centre would incur additional costs for the operator, these can still be compared to the overall net economic benefit from a shift to such a distribution scheme. In the case study analyzed here, all six scenarios involving a city warehouse appear to produce economic benefits in the range of EUR 6 to 13 thousand annually, a quantity that would be sufficient to cover at least a large share of the additional costs. Moreover, as the sensitivity analysis suggests, there are several possible configurations of the distribution system that can lead to even larger economic benefits, e.g., up to EUR 48,100 per year in the case of the maximum for scenario DET2. Nevertheless, there may be an additional strong point in favour of the involvement of local authorities through the provision of no- or low-cost transhipment areas. The potential savings in terms of external costs can be considerable and may exceed the amount of EUR 20,000 annually (Figure 2, scenario DET2 compared to scenario CEZ). Given the potential of the external cost reduction, the choice for local authorities to subsidize at least part of the cost of an intermediate distribution centre can be a cost-effective measure, especially if the installations can be shared among various operators.

There are obviously several additional caveats that should be highlighted. The case study focuses on a specific business application and context. While the beverage sector has similar characteristics in terms of logistic patterns with several other sectors that require deliveries at urban level, the specific delivery sizes, distances and frequencies in each sector may allow different levels of flexibility in terms of vehicle types and fleet mixes. There is considerable uncertainty as regards the assumptions on current and future costs, fuel consumption and emissions, while modelling results cannot be always assumed as accurate. We explored the potential impacts of several of the factors that lead to uncertainty in the sensitivity analysis, but there are still issues that could benefit from a dynamic traffic flow modelling analysis, especially as regards the impact on congestion. The methodology applied is, nevertheless, transparent in terms of the main assumptions and tools used and allows a replication in future research using more adequate assumptions.

5. Conclusions

The case study described here explores the potential for the introduction of green alternatives for last-mile distribution for a specific sector (beverages) in a certain urban area (Seville). The findings for this application can contribute to the overall discussion on the potential of emerging technologies for green distribution and the expected impacts in economic and environmental terms. At the same time, the methodology and the software tool used in this case study can be easily replicated in other sectoral and geographic settings, since they allow a realistic simulation of the potential impact of different approaches to decarbonize the specific distribution system. The results can also be used for the calculation of efficiency indicators in order to benchmark specific sectors or compare across cities.

The analysis presented in this study focuses on the last link in the chain of electric efficiency in cities, specifically the end-use of electric mobility in urban freight transport. As noted in the New Urban Mobility Framework within the EU [2], the complex costs of creating the whole electromobility ecosystem are deliberately not included in the framework. Instead, the framework emphasizes the importance of improving the efficiency and reducing the environmental impacts of urban transport, with a particular focus on the last mile of the delivery. The creation of a comprehensive electromobility ecosystem involves a range of complex costs and challenges, including the development of charging infrastructure, the production of electric vehicles, and the integration of electric mobility into existing transport systems. By focusing on the end-use of electric mobility, our study is able to provide a detailed analysis of the potential benefits and challenges of electrifying urban delivery fleets, without being enmeshed by the complexities of the broader electromobility ecosystem. This approach allows us to identify the key factors that influence the adoption of electric vehicles in urban freight transport, including the cost of vehicles, the availability of charging infrastructure, and the operational characteristics of delivery routes.

Our study provides a detailed analysis of the potential benefits and challenges of electrifying urban delivery fleets, with a focus on the end-use of electric mobility. By recognizing the complexities of the broader electromobility ecosystem but focusing on the last link in the chain of electric efficiency in cities, we are able to provide insights into the ways in which electric vehicles can be used to improve the efficiency and reduce the environmental impacts of urban freight transport. Our findings have implications for policymakers, businesses, and individuals seeking to reduce the environmental impacts of urban transport, and highlight the need for continued research and development into the potential of electric mobility to transform the urban transport sector.

The methodology followed allows the comparison between different alternative technologies using a highly parameterized approach. This is particularly useful in situations of high uncertainty, as is the case of the future evolution of costs and operational characteristics of emerging technologies. The analysis was carried out using current knowledge and estimates, with the possibility of exploring the sensitivity of the results to each particular assumption. A number of robust conclusions can therefore be derived:

• The range and operational characteristics of currently existing electric options are sufficient to fully substitute the conventional options currently used in the case study context. In an urban distribution context, the average total distance covered by distribution vans is lower than 100 km per day, well within the range allowed by electric vans currently available in the market.
• The higher capital costs of electric alternatives can already be compensated by the lower energy costs in most cases. It is expected that the introduction of larger electric freight vehicles (with a payload of 6.5 t or more) would also make that segment competitive with conventional options by 2030.
• The electrification of the vehicles used for distribution would result in clear environmental benefits, especially if the electricity consumed comes from a clean power generation mix. Apart from the direct reductions of CO₂ and pollutant emissions, a distribution system that uses a suitable combination of emerging distribution technologies can also reduce other external costs, such as congestion or noise.
• The findings confirm the potential of electrification in reducing the externalities from urban logistics. In addition, this work contributes to the literature by applying a combined fleet and routing optimization model to consider alternative vehicle types and sizes.
• The average weight and volume of individual shipments may affect the suitability of each alternative. In the specific case study presented here, electric bikes and tricycles can be indeed efficient. Future work can explore applications in different market segments using varying shipment sizes.
• Urban density and the average distribution distance can affect the efficiency of a last-mile delivery system based on electric vehicles. Our results suggest that such an approach is suitable in the case of distributing beverages in Seville. Future work can explore the feasibility of such an approach in areas with a lower density of demand.
• Additional logistic solutions, such as the introduction of intermediate distribution centres, can facilitate the use of smaller and cleaner vehicles for last-mile distribution. In this particular case study, the investment required for the intermediate distribution centre is provided by the city authorities and can be an example of a policy measure that encourages green logistics.